EFFICIENT NETWORK DESIGN FOR ENVIRONMENTAL
MONITORING APPLICATIONS
A Practical Approach
Farooq Sultan and Salam A. Zummo
Electrical Engineering Department, King Fahd University of Petroleum and Minerals, 31261, Dhahran, Saudi Arabia
Keywords: Wireless Sensor Network, LEACH, Network Life Time, Latency.
Abstract: The deployment of a wireless sensor network is highly dependent on the target environment. Once the
characteristics of the desired area are know, the question of network size arises. Factors like transmission
power level, cost of network deployment and the coverage area directly affect the size of the network. This
paper analyzes the behaviour of a typical multi-hop wireless sensor network operating in an outdoor
environment. By considering two separate cases; fixed cost and fixed deployment area, we present best
network set-up statistics based on actual received power measurements.
1 INTRODUCTION
The concept of a wireless sensor network (WSN)
involves the integration of sensing, processing and
communication abilities to create highly autonomous
networks. In order to combine sensors, processors
and radio devices, a detailed study of the desired
application as well as the capabilities of the
available hardware has to be done. Wireless nodes
are the building blocks of WSNs; sensing,
processing and communicating abilities are
integrated to produce miniature devices that can
form and maintain a network. The origins of WSN
trace back to the distributed networks program
(Lacoss, 1986) launched by DARPA in the late 80s
using bulky wireless devices to convey information
to the end-user. The advancement in the CMOS
technology has led to extremely small devices with
exceptionably high processing abilities giving rise to
tiny wireless nodes (MicaZ, Telos etc. by Crossbow
Technologies).
The set-up and operation of a WSN is highly
dependent on the application. The main emphasis of
the work presented in this paper is towards
environmental monitoring scenarios. Environmental
data collection requires the collection of multiple
sensor readings from geographically different
locations over time. These readings need to be
transmitted to a base station, where analysis can be
done to deduce results (Sun, 2010). From the
network deployment perspective, a large number of
nodes need to be placed (randomly or
deterministically) having the ability to relay/forward
information to the sink node.
This paper presents a model for the efficient
deployment of a WSN in an outdoor environment.
By carrying out received power measurement
analysis for a specially designed general purpose
wireless sensor node in an outdoor environment,
simulations have been done for the cases of fixed
cost and fixed coverage area. The designed
application provides the required network size, the
transmission power requirement and the coverage
area statistics for different combinations of the input
parameters.
This paper is organized as follows; Section 2
presents an overview of the WSN architecture,
Section 3 has the procedure for received signal
strength measurements, Results and analysis is
presented in Section 4 and the paper concludes with
a summary in Section 5.
2 WSN ARCHITECTURE
2.1 Setup Requirements
Environmental monitoring usually requires periodic
data logging, giving rise to extremely low data rates.
Compared to traditional single hop networks, the
WSN has some unique features (Ye, 2009);
69
Sultan F. and A. Zummo S..
EFFICIENT NETWORK DESIGN FOR ENVIRONMENTAL MONITORING APPLICATIONS - A Practical Approach.
DOI: 10.5220/0003614900690072
In Proceedings of the International Conference on Wireless Information Networks and Systems (WINSYS-2011), pages 69-72
ISBN: 978-989-8425-73-7
Copyright
c
2011 SCITEPRESS (Science and Technology Publications, Lda.)
1) All nodes can form a network based on
some protocol.
2) All nodes can sense and route the
information to nearby nodes.
3) Nodes can join or leave the network
without any management issues.
The above mentioned requirements enable the
nodes to operate in a highly collaborative manner
thereby preserving the energy supplies for the longer
operation.
2.2 Network Structure
As mentioned in Section 2.1, the network topology
of the WSN is governed by the protocols adopted by
the nodes. For the results presented in this paper, a
hierarchical network topology is assumed based on
the low energy adaptive clustering hierarchy
(LEACH) (Baghyalakshmi, 2009). According to this
approach, all the nodes in the network consider
themselves as either cluster heads (CH) or member
nodes (MN). Figure 1 shows a graphical
representation of such a network.
Figure 1: Network layout of a typical WSN for
environmental monitoring.
The CH has the responsibility of collecting the
data from its MNs and relaying it to the next-hop
CH until it reaches the base station. MNs only
communicate with their respective CH using the
least possible transmit power (P
t
), whereas CH can
adjust their transmit power levels based on the
requirement of the network topology.
3 RECEIVED SIGNAL
STRENGTH (RSS)
MEASUREMENTS
RSS measurements were carried out using the
custom designed wireless sensor node in an open
outdoor environment (University stadium). The
transceiver unit, CC2420 (Chipcon, 2005), used on
the used wireless sensor node, offers five different P
t
settings including 0dBm. Moreover, the CC2420 has
the ability to calculate the RSS of the received
packet fairly accurately.
The actual RSS of a packet is not considered to
be an accurate representation of the environmental
conditions for large WSNs deployments as packets
can be received with extremely low RSS (<90 dBm).
Instead a different metric, packet receive rate (PRR),
is taken into account to represent the fraction of
packets received. PRR enables a much closer
understanding of the channel conditions by
considering packet loss; which is vital in
environmental monitoring applications.
Figure 2: PRR with varying transmitter-receiver separation
for four different P
t
in an outdoor environment.
Measurements for PRR relative to the
transmitter-receiver separation were done for four P
t
levels as shown in Figure 2. For 0 dBm, the
measurements were inconclusive as a perfect
reception was achieved even up to a separation of 70
m. The data presented in Figure 2 was used as a base
for determining the signal quality at a given P
t
. All
the statistics presented in the following sections of
the paper are based on these measurements.
4 RESULTS AND ANALYSIS
4.1 Fixed Cost and PRR
For this case, the cost of the network, in the number
of nodes available, and the PRR are provided as the
input to the model. This case presents the situation
in which, the user has a fixed number of nodes and
he requires a certain PRR as a measure of the QoS.
Figure 3 shows the results of a network comprising
of 50 nodes operating at a PRR of 0.9.
WINSYS 2011 - International Conference on Wireless Information Networks and Systems
70
Figure 3: Network life (in hours) for different cluster sizes
for a network comprising of 50 nodes operating at a PRR
of 0.9 for an outdoor vicinity of 50 m
2
.
From Figure 3, it can be observed that, for a
fixed number of nodes, regardless of the transmit
power level used; a lightly packed cluster topology
results in the maximum possible life of the network.
When the deployment area is also fixed, this
condition implies a larger number of lightly packed
clusters operating fairly close to each other. Due to
less inter-cluster distance, low P
t
is required to fulfil
the communication requirements between adjacent
clusters, hence increasing the overall operating life
of the WSN. As the individual clusters become
populated, the separation between them must be
increased to ensure maximum coverage of the given
area, thereby increasing the requirement of P
t,
which
results in lowering the network operating time.
Tables 1 and 2 present the network lifetime
statistics for the case presented in Figure 3 for two
different PRR requirements. The PRR is a measure
of the link quality and is not dependent on the
network life time as observed from the figures in
Table 1 and Table 2. The maximum inter-cluster
spacing is inversely proportional to the PRR and is
extracted from the measurement data presented in
Figure 2. As the desired PRR is decreased (0.5 for
worst case), the inter cluster spacing increases
enabling the network to span a much larger area as
compared to a PRR of 0.9.
4.1.1 End-to-End Latency Analysis
The total time required for a packet from its
generation to the delivery at the base station is
referred to as the end-to-end latency. For
environmental monitoring applications in critical
situations (hospitals etc.), the latency plays an
important role and must be minimized for the data to
be of any meaningful value. NS-2 simulations of the
network mentioned in Figure 3, give the stats
Table 1: Network life (hours) and corresponding
maximum inter-cluster separation for PRR=0.9.
Cluster Size
(nodes)
P
t
(dBm)
-5 -10 -15 -25
2 53 56 59 61
5 32 35.13 37 38.3
10 30 31 33 34
25 28 29 31 32
50 27 29 30 31
Inter-cluster
Spacing (m)
12.6 2.5 1.3 0.6
Table 2: Network life (hours) and corresponding
maximum inter-cluster separation for PRR=0.5.
Cluster Size
(
nodes
)
-5 -10 -15 -25
2 53 56 59 61
5 32 35.13 37 38.3
10 30 31 33 34
25 28 29 31 32
50 27 29 30 31
Inter-cluster
Spacing (m)
23.7 6.7 3.3 0.9
provided in Figure 4.
When small cluster sizes are adopted, a large
number of clusters are required to cover a given
region. Although this increases the network
operation life, but makes the packet reach the base-
station over multiple hops. Delay at individual hops
adds up and becomes significant due to the sleep
cycles and congestion conditions.
Figure 4: End-to-end latency for a 50 node network
operating at varying cluster sizes at a PRR of 0.9 for an
outdoor vicinity of 50 m
2
.
0 5 10 15 20 25 30 35 40 45 50
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
5.5
Cluster size (nodes)
End-to-end latency (s)
EFFICIENT NETWORK DESIGN FOR ENVIRONMENTAL MONITORING APPLICATIONS - A Practical Approach
71
As it can be observed from Figure 4, the latency
could be as high as 5.2 seconds.
4.2 Fixed Area and PRR
Practical situations might require the user to specify
the coverage area instead of the cost (as in Section
4.1). For a PRR of 0.9 and an area of 50 m
2
the
network size (in clusters) is presented in Figure 5.
Figure 5: Network size (clusters) at different inter-cluster
transmission powers with a PRR of 0.9 for an outdoor
vicinity of 50 m
2
.
As expected, for a low P
t
a larger number of
clusters are required. The number of nodes within a
cluster is dependent on the actual intra-cluster
transmission power levels selected which depends
on the implementation scenario.
5 CONCLUSIONS
WSN deployment is a tricky job unless handled
wisely. This paper presented a model based on
actual PRR measurements to study the relationship
between critical factors like network lifetime, end-
to-end latency and the coverage plan for a given
area. By analyzing the simulation results, a trade-off
between network life and the end-to-end delay was
observed. Based on the target environment and the
required parameters, this model tends to provide the
best deployment plan according to the desired PRR.
ACKNOWLEDGEMENTS
The authors would like to acknowledge King Fahd
University of Petroleum and Minerals (K.F.U.P.M.)
for providing a highly conducive environment for
research. In addition, we would like to acknowledge
King Abdulaziz City of Science and Technology
(KACST) for funding this project under project
number ط م6 -2 .
REFERENCES
Lacoss, R. 1986. Distributed sensor networks. In
Technical Report, DARPA Project.
Sun, D., Jiang, S., Wang, W., Tang, J., 2010. Wireless
sensor network design and implementation in a tea
plantation for drought monitoring. In 2010
International Conference on Cyber-Enabled
Distributed Computing and Knowledge Discovery.
Ye, D., Gong, D., Wang, W., 2009. Application of
wireless sensor networks in environmental
monitoring.. In Second International Conference on
Power Electronics and Intelligent Transportation
Systems..
Baghyalakshmi, D., Ebenezer, J., SatyamurtY, S. 2010.
Low latency and energy-efficient routing protocols for
wireless sensor networks. In International Conference
on Wireless Communications and Sensor computing.
2.4 GHz IEEE 802.15.4/ZigBee-ready RF Transceiver.
2005. Chipcon Products from Texas Instruments,
CC2420.
-25 -20 -15 -10 -5
0
5
10
15
20
25
30
35
40
45
Inter cluster transmit power (dBm)
Network size (clusters)
WINSYS 2011 - International Conference on Wireless Information Networks and Systems
72