Performance Evaluation of Wireless Sensor Networks for Different Radio
Models Considering Mobile Event
Tao Yang†, Makoto Ikeda†, Leonard Barolli ‡, Arjan Durresi ††, Fatos Xhafa †‡
†Graduate School of Engineering
Fukuoka Institute of Technology (FIT)
3-30-1 Wajiro-Higashi, Higashi-Ku, Fukuoka 811-0295, Japan
E-mail: {bd07003, bd07001}@bene.fit.ac.jp
‡Department of Information and Communication Engineering
Fukuoka Institute of Technology (FIT)
3-30-1 Wajiro-Higashi, Higashi-Ku, Fukuoka 811-0295, Japan
E-mail: barolli@fit.ac.jp
††Department of Computer and Information Science
Indiana University Purdue University at Indianapolis (IUPUI)
723 W. Michigan Street SL 280 Indianapolis, IN 46202, USA
E-mail: durresi@cs.iupui.edu
†‡Department of Languages and Informatics Systems
Technical University of Catalonia
C/Jordi Girona 1-3, 08034 Barcelona, Spain
E-mail: fatos@lsi.upc.edu
Abstract
In this paper, we consider the behavior of a wireless sen-
sor network for TwoRayGround and Shadowing radio mod-
els for the case of mobile event. By means of simulations,
we analyse the performance of AODV protocol. In the pre-
vious work, we considered that the event node is stationary
in the observation field. In this work, we want to inves-
tigate how the sensor network performs in the case when
the event node moves. The simulation results show that the
shadowing phenomena, by destroying the regularity of the
network, reduce the mean distance among nodes and at the
same time increase the interference level and the latency of
packet transmission. We found that for mobile event, the
Goodput and routing efficiency of TwoRayGround is better
than Shadowing, but the Depletion of Shadowing is better
than TwoRayGround.
1. Introduction
In recent years, technological advances have lead to
the emergence of distributed Wireless Sensor Networks
(WSNs) which are capable of observing the physical world,
processing the data, making decisions based on the ob-
servations and performing appropriate actions. These net-
works can be an integral part of systems such as battle-field
surveillance and microclimate control in buildings, nuclear,
biological and chemical attack detection, home automation
and environmental monitoring [1, 2].
Wireless sensor network simulation is an important part
of the current research. A large number of algorithms were
first implemented and evaluated using several network sim-
ulators like ns-2. Most MANET routing protocols have
been developed and tested in that fashion, and only later
they evolved towards real world implementations.
Recently, there are many research works for sensor net-
works [3, 4]. In our previous work [5], we implemented
a simulation system for sensor networks consider different
2010 International Conference on Complex, Intelligent and Software Intensive Systems
978-0-7695-3967-6/10 $26.00 © 2010 IEEE
DOI 10.1109/CISIS.2010.91
180
protocols ( e.g: AODV, DSR, DSDV, OLSR. ) and different
propagation radio models. In [6], we analysised the perfor-
mance of the WSNs considering different topologies with
the irregular radio model. Also, we analysised the perfor-
mance of our proposed simulation systems. But, we con-
sidered that the event node is stationary in the observation
field. However, in many applications the event node may
move. For example, in an ecology environment the animals
may move randomly. Another example is when an event
happens in a robot or in a car.
In this work, we want to investigate how the sensor net-
work performs in the case when the event moves and use
also Shadowing propagation model. We carry out simula-
tions for lattice topology, TwoRayGround and Shadowing
radio model considering Ad-hoc On-demand Distance Vec-
tor (AODV) protocol.We compare the simulation results for
the mobile event using different propagation radio model.
The simulation results have shown that the goodput for the
case of mobile event node is better than the stationary event
node using AODV protocol and TwoRayGround model.
Also, the goodput for the mobile event node case does not
change too much compared with the stationary event case
using AODV and Shadowing model, but the goodput is not
good when the number of nodes is increased.
The paper is organized as follows. In Sections 2, we ex-
plain the proposed network simulation model, simulation
topology, routing protocols and radio models. In Secton 3,
we introduce the event detection and transport. In Section
4, we introduce the goodput, Routing Efficiency (RE) and
energy depletion concepts. In Section 5 we present simula-
tion results. Finally, conclusions of the paper are given in
Section 6.
2. Proposed Network Simulation Model
Proposed network simulation model is shown in Fig. 1.
In our WSNs model, every node detects the physical phe-
nomenon and sends back to the sink node data packets. We
suppose that the sink node is more powerful than sensor
nodes and it is always located at the borders of the service
area. This model can be used for remote monitoring of haz-
ard or inaccessible areas [7]. We analyse the performance of
the network in a fixed time interval. This can be considered
as the available time for the detection of the phenomenon
and its value is application dependent. In this paper, we con-
sider that a mobile event is moving randomly in the WSNs
field. In Fig. 2 is shown one pattern of movement event’s
path. We implemented a simulation system for WSNs con-
sidering moving event using ns-2. We evaluated the good-
put, routing efficiency and consumed energy of AODV pro-
tocol for TwoRayGround and Shadowing propagation mod-
els in case of the lattice topology.
Modeling
Topology
Radio Model
Routing Protocol
Physical & MAC
CSMA
Regular
Random
Shadowing
TwoRayGround
AODV, DSR
OLSR, Others
Simulation System
Protocol Evaluation
Data For Real
Applications
IEEE802.11 Zigbee
Sensor Network
Network Stack
Figure 1. Proposed network simulation
model.
2.1. Topology
For the physical layout of the WSNs, two types of de-
ployment has been studied so far: the random and the lat-
tice deployment. In the former, nodes are supposed to be
randomly distributed, while in the latter nodes are vertexes
of particular geometric shape, e.g. a square grid. For space
constraints, we present results for the square grid topology
only. In this case, in order to guarantee the connectedness
of the network we should set the transmission range of ev-
ery node to the step size, d, which is the minimum distance
between two rows (or columns) of the grid. In fact, by this
way the number of links that every node can establish, a.k.a
the node degree is D =4. By using Cooper’s theorem [8]
along with some power control techniques, one could use
also D =2
1
. However, we assume all nodes to be equal
and then the degree is fixed to 4. Nodes at the borders have
D =2.
2.2. Routing Protocol
We are aware of many proposals of routing protocols
for ad-hoc networks during recent years. Here, we con-
sider AODV protocol [9]. The AODV is an improvement
of DSDV to on-demand scheme. It minimize the broadcast
packet by creating route only when needed. Every node in
network should maintain route information table and par-
ticipate in routing table exchange. When source node wants
1
By using the theorem in [8], we can say that a simple 2 regular network
is almost surely strongly 2 connected.
181
Sensor node Event Sink
Figure 2. One pattern of event movement
path.
to send data to the destination node, it first initiates route
discovery process. In this process, source node broadcasts
Route Request (RREQ) packet to its neighbors. Neighbor
nodes which receive RREQ forward the packet to its neigh-
bor nodes. Neighbor nodes which receive RREQ forward
the packet to its neighbors, and so on. This process con-
tinues until RREQ reach to the destination or the node who
know the path to destination. When the intermediate nodes
receive RREQ, they record in their tables the address of
neighbors, thereby establishing a reverse path. When the
node which knows the path to destination or destination
node itself receive RREQ, it send back Route Reply (RREP)
packet to source node. This RREP packet is transmitted by
using reverse path. When the source node receives RREP
packet, it can know the path to destination node and it stores
the discovered path information in its route table. That is
the end of route discovery process. Then, AODV performs
route maintenance process. In route maintenance process,
each node periodically transmits a Hello message to detect
link breakage.
2.3. Propagation Radio Model
In order to simulate the detection of a natural event, we
used the libraries from Naval Research Laboratory (NRL)
[10]. In this framework, a phenomenon is modeled as a
wireless mobile node. The phenomenon node broadcasts
packets with a tunable synchrony or pulse rate, which repre-
sents the period of occurrence of a generic event
2
. These li-
2
As a consequence, this model is for discrete events. By setting a suit-
able value for the pulse rate, it is possible in turn to simulate the continuous
braries provide the sensor node with an alarm variable. The
alarm variable is a timer variable. It turns off the sensor if
no event is sensed within an alarm interval. In addition to
the sensing capabilities, every sensor can establish a multi-
hop communication towards the Monitoring Node (MN) by
means of a particular routing protocol. This case is the op-
posite of the polling scheme.
Although not optimal for multi-hops WSNs, we assume
that the MAC protocol is the IEEE 802.11 standard. This
serves to us as a baseline of comparison for other con-
tention resolution protocols. The receiver of every sensor
node is supposed to receive correctly data bits if the received
power exceeds the receiver threshold, γ. This threshold de-
pends on the hardware
3
. As reference, we select param-
eters values according to the features of a commercial de-
vice (MICA2 OEM). In particular, for this device, we found
that for a carrier frequency of f = 916MHz and a data
rate of 34KBaud, we have a threshold (or receiver sensitiv-
ity) γ|
dB
= −118dBm [11]. The calculation of the phe-
nomenon range is not yet optimized and the phenomenon
propagation is assumed to follow the propagation laws of
the radio signals. Table 1 shows some typical values of
Shadowing deviation. In Fig. 3 and Fig. 4 are shown the
transimisson range of TwoRayGround and Shadowing mod-
els [12]. TwoRayGround model considers both the direct
path and a ground reflection path. It is applied in the envi-
ronments which are like plains and have no obstacles. How-
ever, the transmission range of Shadowing model is random.
This model is applied in the environments which have ob-
stacles and are hardly to transmit data directly. In particular,
the emitted power of the phenomenon is calculated accord-
ing to a TwoRayGround propagation model. The received
power at distance d is predicted by:
P
r
(d)=
P
t
G
t
G
r
h
2
t
h
2
r
d
4
L
(1)
where G
t
and G
r
are the antenna gains of the transmitter
and the receiver, h
t
and h
r
are the heights of the transmit
and receive antennas respectively, and L (L ≥ 1) is the sys-
tem loss.
The Shadowing model assumes that the received power
at the sensor node is:
P
r
(d)|
dB
= P
t
|
dB
− β
0
− 10α log
d
d
0
deterministic part
+ S
dB
random part
(2)
where β
0
is a constant. The term S
dB
is a random variable,
which accounts for random variations of the path loss. This
signal detection such as temperature or pressure.
3
Other MAC factors affect the reception process, for example the Car-
rier Sensing Threshold (CST) and Capture Threshold (CP) of IEEE.802.11
used in ns-2.
182
20
40
60
80
100
30
210
60
240
90
270
120
300
150
330
180 0
Case of deterministic model (σ
2
=0)
Figure 3. Transimisson range of TwoRay-
Ground model.
50
100
150
200
250
30
210
60
240
90
270
120
300
150
330
180 0
Case of random model, σ
2
=2dB
Figure 4. Transimisson range of Shadowing
model.
variable is also known as log-normal shadowing, because it
is supposed to be Gaussian distributed with zero mean and
variance σ
2
dB
,thatisS
dB
∼N(0,σ
2
dB
). Given two nodes,
if P
r
>γ,whereγ is the hardware-dependent threshold,
the link can be established. The case of σ =0, α =4,
d>d
0
is the TwoRaysGround model. In Shadowing model
in addition to the direct ray from the transmitter towards the
receiver node, a ground reflected signal is supposed to be
present. Accordingly, the received power depends also on
Table 1. Some typical values of Shadowing
deviation S
dB
.
Environment S
dB
Outdoor 4 to 12
Office, Hard partition 7
Office, Soft partition 9.6
Factory, Line-of-sight 3 to 6
Factory, Obstructed 6.8
Table 2. Topology settings.
Lattice
Step d =
L
√
N−1
m
Service Area Size L
2
= (800x800)m
2
Number of Nodes N =12, 64, 100, 256
Transmission Range r
0
= d
the antenna heights and the pathloss is:
β =10log
(4πd)
4
L
G
t
G
r
h
t
h
r
λ
2
(3)
Energy Model The energy model concerns the dy-
namics of energy consumption of the sensor. A widely used
model is as follows [13]. When the sensor transmits k bits,
the radio circuitry consumes an energy of kP
Tx
T
B
,where
P
Tx
is the power required to transmit a bit which lasts T
B
seconds. By adding the radiated power P
t
(d),wehave:
E
Tx
(k, d)=kT
B
(P
Tx
+ P
t
(d)) .
Since packet reception consumes energy, by following the
same reasoning, we have:
E(k, d)=E
Tx
(k, d)+E
Rx
(k, d)=kP
Tx
T
B
+ kT
B
P
t
(d)
+ kP
Rx
T
B
(4)
where P
Rx
is the power required to correctly receive (de-
modulate and decode) one bit. In Tables 2 and 3, we sum-
marise the values of parameters used in our WSNs. Let us
note that the power values concern the power required to
transmit and receive one bit, respectively. They do not refer
to the radiated power at all. This is also the energy model
implemented in the widely used ns-2 simulator. It is take a
long time to simulate a simulation system with large num-
ber of nodes. So, we investgated the performance of WSNs
with the number of nodes shown in Table 2.
183
Table 3. Radio model and system parameters.
Radio model parameters
Path Loss Coefficient α =2.7
Variance σ
2
dB
= 16dB
Carrier Frequency 916MHz
Antenna omni
Threshold (Sensitivity) γ = −118dB
Other parameters
Reporting Frequency T
r
=[0.1, 1000]pps
1
Interface Queue Size 50 packets
UDP Packet Size 100 bytes
Detection Interval τ 30s
1
packet per seconds
Interference In general, in every wireless network the
electromagnetic interference of neighboring nodes is al-
ways present. The interference power decreases the Signal-
to-Noise-Ratio (SNR) at the intended receiver, which will
perceives a lower bit and/or packet error probability. Given
a particular node, the interference power depends on how
many transmitters are transmitting at the same time of the
transmission of the given node. In a WSNs, since the num-
ber of concurrent transmissions is low because of the low
duty-cycle of sensors, we can neglect the interference. In
other words, if we define duty-cycle as the fraction between
the total time of all transmissions of sensor data and the total
operational time of the net, we get always a value less than
0.5. In fact, the load of each sensor is 1 because sensors
transmit data only when an event is detected [13]. How-
ever, it is intuitive that in a more realistic scenario, where
many phenomena trigger many events, the traffic load can
be higher, and then the interference will worsen the perfor-
mance with respect to that we study here. Consequently,
we can fairly say that the results we get here should be con-
sidered as an upper bound on the system performance with
respect to more realistic scenarios.
3 Event Detection and Transport
Here, we use the data-centric model similar to [14],
where the end-to-end reliability is transformed into a
bounded signal distortion concept. In this model, after sens-
ing an event, every sensor node sends sensed data towards
the MN. The transport used is a UDP-like transport, i.e.
there is not any guarantee on the delivery of the data. While
this approach reduces the complexity of the transport pro-
tocol and well fit the energy and computational constraints
of sensor nodes, the event-reliability can be guaranteed to
some extent because of the spatial redundancy. The sensor
T
0
r
f()
T
r
WSN
Target event−reliability
Event−reliability
Figure 5. Representation of the transport
based on the event-reliability.
node transmits data packets reporting the details of the de-
tected event at a certain transmission rate
4
. The setting of
this parameter, T
r
, depends on several factors, as the quan-
tization step of sensors, the type of phenomenon, and the
desired level of distortion perceived at the MN. In [14], the
authors used this T
r
as a control parameter of the overall
system. For example, if we refer to event-reliability as the
minimum number of packets required at MN in order to re-
liably detect the event, then whenever the MN receives a
number of packets less than the event-reliability, it can in-
struct sensor nodes to use a higher T
r
. This instruction is
piggy-backed in dedicated packets from the MN. This sys-
tem can be considered as a control system, as shown in
Fig. 5, with the target event-reliability as input variable and
the actual event-reliability as output parameter. The target
event-reliability is transformed into an initial T
0
r
. The con-
trol loop has the output event-reliability as input, and on
the basis of a particular non-linear function f (·), T
r
is ac-
cordingly changed. We do not implement the entire control
system, but only a simplified version of it. For instance, we
vary T
r
and observe the behavior of the system in terms of
the mean number of received packets. In other words, we
open the control loop and analyze the forward chain only.
4. Goodput, Routing Efficiency and Consumed
Energy
In this section, we present the simulation results of our
proposed WSNs. We simulated the network by means of ns-
2 simulator, with the support of NRL libraries
5
. The Good-
put is defined at the sink, and it is the received packet rate
divided by the sent packets rate. Thus:
G(τ)=
N
r
(τ)
N
s
(τ)
(5)
4
Note that in the case of discrete event, this scheme is a simple packet
repetition scheme.
5
Since the number of scheduler events within a simulated WSNs can
be very high, we applied a patch against the scheduler module of ns-2 in
order to speed up the simulation time [15].
184