Performance Evaluation of a Wireless Sensor Network Based Tracking System
Nadeem Ahmed, Yifei Dong,
Salil S Kanhere, Sanjay Jha
School of Computer Science and Engineering,
The University of New South Wales,
Sydney, Australia.
{nahmed, ydon, salilk, sanjay}
@cse.unsw.edu.au
Abstract
In this paper, we present a comprehensive analysis of
the performance of a wireless sensor network based tar-
get tracking system using the Particle Filter. In particular,
we evaluate the effect ofvarious network design parameters
such as the number ofnodes, number ofgenerated particles,
and sampling interval on the tracking accuracy and compu-
tation time of the tracking system. Based on our analysis,
we also present recommendations on suitable values for the
relevant network design parameters, which provide a rea-
sonable tradeoff between accuracy and computational ex-
pense for this problem. In addition, we also analyse the
theoretical Cramer-Rao Bound as the benchmark for the
best possible tracking performance. We demonstrate that
the results from our simulations closely match the theoreti-
cal bounds.
We also present initial results from experiments compris-
ing of a 25 node wireless sensor network. Initial experi-
mental results are promising and show that the PF based
estimation is suitable for detection and tracking using inex-
pensive wireless sensor network devices.
1 Introduction
Wireless Sensor Networks (WSN) are increasingly used
in a variety of applications ranging from environmental
monitoring to industrial automation. A particularly promis-
ing military application involves using WSN for detecting
and tracking moving targets such as tanks, vehicles and
troops. Detection and tracking of targets is a mature and
well-established research area. However, the proposed so-
lutions rely on expensive and bulky sensors. Using low-cost
sensor nodes is an attractive and complementary approach.
Among several tracking algorithms in literature, for non-
linear filtering the Particle Filter (PF) [1], has been a pop-
ular choice. The PF is also known as the sequential Monte
978-1-4244-2575-4/08/$20.00@2008 IEEE
Authorized licensed use limited to: IEEE Xplore. Downloaded on October 30, 2008 at 00:59 from IEEE Xplore. Restrictions apply.
Mark Rutten, Travis Bessell,
Neil Gordon
ISR Division,
Defence Science and Technology Organisation,
Edinburgh, Australia.
{Mark.Rutten, Travis.Bessell, Neil.Gordon}
@dsto.defence.gov.au
Carlo method, because it approximates a belief state for
the presence of the target by means of many but finite ran-
dom samples. In the past [2-4] PF was shown to be a suit-
able candidate for deployment in tracking applications us-
ing WSN. However, the achievable tracking and detection
performance of the system, is significantly influenced by
several network design parameters, namely: the number of
deployed sensor nodes, the number of generated particles,
and the sampling interval ofsensor nodes (which is defined
as the idle interval between successive samples). All prior
work simply assumes a particular set of values for these pa-
rameters without providing insight into how they affect the
behaviour of the tracking system. To the best of our knowl-
edge, this is the first study that provides a detailed and thor-
ough examination of the effects of these parameters on the
tracking and detection performance in a realistic simulation
environment. Based on our observations, we also suggest
suitable values for the relevant network design parameters.
In our evaluations, we consider the problem of simulta-
neous detection and tracking of an object moving through a
particular target region, where a WSN has been deployed.
The sensor nodes are equipped with acoustic sensors, which
periodically sample the ambient sound and relay the mea-
sured samples to a central base station. The base station ex-
ecutes the PF algorithm on the collected data-set. We used
the Ns2 discrete event simulator [5] for this study. The two
metrics that we have used to evaluate the performance of
the system are: estimation of the tracking error, Le., the
distance between the predicted and the actual location of
the target and computation time Le., the time required by
the PF to execute and detect the target. We evaluated the
effect of each of the aforementioned network parameters on
these two performance metrics.
We have also analysed the theoretical Cramer-Rao
Bound [6] as the benchmark for the best possible track-
ing performance and compared the results of our simulated
PF algorithm with respect to this bound. Our compari-
son shows that the accuracy of the simulated PF algorithm
163
closely matches the theoretical bound.
We have conducted experiments with off-the-shelf WSN
devices (Xbow MicaZ motes) to verify the validity of our
simulation results. We first conducted calibration exper-
iments to characterize the gain for each sensor and noise
variations in a realistic environment. These calibration val-
ues are then used by the PF algorithm for tracking purposes.
The experimental results are promising and the PF based
tracking system is able to track the movement of the target
with high accuracy.
The rest of this paper is organised as follows: Section 2
summarises related work. In Section 3 we provide a brief
overview of the PF and discuss how we adapt it to our sys-
tem. Interested readers are referred to a book by one of
the co-authors [1] for further elaboration. The simulation-
based evaluations along with relevant discussions on appro-
priate system design parameters are presented in Section 4.
Section 5 derives the Cramer-Rao bound and presents a
comparison of our simulation results with this theoretical
bound. Section 6 presents results from our initial experi-
ments using a 25 node WSN. Finally, Section 7 concludes
the paper and discusses future work.
2 Related Work
In this section, we present a brief overview of previous
work on using miniature sensor devices for target tracking.
The work introduced in [4], is one of the earliest attempts to
use tiny acoustic sensor devices for tracking purposes. The
target is estimated via triangulation, i.e., comparing the dif-
ference in sound propagation delays from the sound source
to different acoustic sensors. In [2], Gu et a!. developed a
light-weight multi-modal detection algorithm for Mote level
micro-sensors. They found out that simple fusion algo-
rithms such as moving averages with thresholds are useful
in object detection using WSN. Unfortunately, both of these
studies assume that the sensor readings are free of ambient
noise, which is a highly unrealistic assumption. In a typical
outdoor environment (especially given the hostile nature of
battlefields), it is expected that the sensor readings would
be significantly influenced by ambient noise. In addition,
given that the sensor nodes are small form factor devices
and low-cost, it is expected that the readings would inher-
ently be noisy due to their low fidelity.
In [3], Duarte et a!., evaluated different machine learn-
ing algorithms in the context of vehicle detection. The au-
thors proposed a two level detection architecture to increase
the reliably of the PF. Different target detection algorithms,
such as K-nearest neighbour, maximum likelihood classi-
fier and support vector machine classifier, were evaluated at
a local node level. Then, the results of the local node level
evaluation are passed to a group, which is formed dynam-
ically. The fusion algorithms are performed at the group
level nodes. However, resource-intensive tasks such as Fast
Fourier Transform (FFf) are required to be performed at
164
Authorized licensed use limited to: IEEE Xplore. Downloaded on October 30, 2008 at 00:59 from IEEE Xplore. Restrictions apply.
the local level nodes. Therefore, it is not suited for low-cost
sensors.
Simon et a!., designed and implemented a sniper local-
isation system based on acoustic signal processing and tri-
angulation in [7]. Special hardware (Digital Signal Process-
ing board) was designed in [7] for the resource-intensive
acoustic signal processing tasks. In [8], He et a!., designed
and implemented a WSN with magnetic, acoustic, and mo-
tion sensors, which could classify a moving target such as
a walking person or a vehicle. The motion sensor used in
this work is a micro-power impulse radar. Due to their high
cost (typically US$5K), they may not be a suitable choice
for many WSN systems.
Coates et a!., also make use of the PF for target tracking
in [9]. However, they propose to model each mote as a par-
ticle. Consequently, the corresponding real-world deploy-
ment would require thousands of motes to achieve tracking
performance comparable to their simulation results. Fur-
ther, the authors assumed that the motes have a fixed sens-
ing range of 8 meters, a fixed detection probability of 0.7
within this sensing range and also assumed absence of any
communication errors. All these assumptions make it diffi-
cult to apply the proposed algorithm in a real-world system.
What is lacking is a comprehensive study of a WSN-
based detection and tracking system, based on realistic as-
sumptions, that explains the effect of various network pa-
rameters on the tracking accuracy of the system. This is
precisely the objective of our work.
3 Overview of the Particle Filter
In this section we describe a recursive Bayesian tracking
algorithm referred to as a Track-Before-Detect Particle Fil-
ter (TBD-PF) in [1]. Using this type of filter allows the in-
formation in the measurements from all sensors to be incor-
porated exactly into the estimation of the target state. The
PF is a suboptimal non-linear filter that performs estima-
tion using sequential Monte Carlo methods to represent the
probability density function of the target state. There are
several advantages of using a PF based estimator over other
non-linear filtering approaches including
• Target presence and absence are explicitly modelled by
the probability function.
• The method can track targets moving randomly in the
field of deployment.
• Non-Gaussian noise in sensor readings can be incorpo-
rated into the filter by estimating the distribution func-
tion of this noise. This incorporates the noise due to
calibration errors in sensors in addition to the environ-
mental noise.
• It permits us to detect targets with variable levels of
intensity.
The design of the filter has been limited to estimating the
state of a single target for the purposes of this work.
We now give a step by step overview of the operation of
the PF and discuss how it is adapted to meet the require-
ments of our system.
3.1 Target Model
We begin by assuming that N sensors are deployed in
a n x m surveillance area, with known positions (Xi, Yi),
i E {I ... N}. Suppose that a target is moving in this area
according to a known dynamic model:
X t +1 = FXt + Vt
where
where ql and q2 control the uncertainty of the target trajec-
tory in position and intensity respectively.
3.2 Sensor Model
Each sensor i E {I ... N} provides a measurement
(acoustic in our case) at discrete instants of time t. The
source is assumed to be emitting a white acoustic signal of
constant power. This measurement is made by calculating
the variance (the acoustic power) of 1000 acoustic samples
taken at a constant sampling rate. It is assumed that each of
these samples is approximately Gaussian distributed, with
zero mean and variance given by
(1)
ai2 + "Ii I t d-t,i'e (7)

• Vt is the process noise, normally assumed to be Gaus- where It is the intensity of the target at time t, is the a;
sian noise with covariance matrix Q. measurement noise variance for sensor i, "Ii is the gain fac-
tor for sensor i and ~ is the path loss, assumed identical for
• t is time index.
each sensor. The distance dt,i from the target to sensor i at
• X t is the target state vector defined as follow: time t is given by
Xt = [Xt, Xt, Yt, 1h, It]T (2) dt,i = J(Xi - Xt)2 + (Yi - Yt)2. (8)
here (Xt, Yt) and (Xt,1it) denote the position and the It can be easily shown that the resulting measurement is X2 -
velocity of the target and It denotes the target's inten- distributed with 1000 degrees of freedom. This distribution
sity. can be quite accurately approximated by a Gaussian with
The existence of the target in the data is also modelled. mean and variance
The target existence variable, E t , can take on two values,
namely E t = 0 indicating the absence of the target and hi(Xt ) = a; + "Iiltd~f (9)
E t = 1 denoting its presence. The target can appear at any
place and at any time-step. Following its appearance, the
Ri(Xt ) = ~s c [0-; + 'Yiltd~fr (10)

target proceeds on a trajectory until it disappears, i.e., the where c is a constant used to model nonlinearities in the
intensity of the target signal strength falls below the sensor's system and N s is the number of samples in the measurement
sensitivity level. We can model the transitional probabilities (1000 in this case). Section 4.1 discusses the method of
of the target birth, P b , and its death, Pd , as follows calibration used to determine a;
and "Ii for each sensor.
Pb = Pr{Et = llEt - 1 = O} (3) The sensor model can thus be summarised as
Pd = Pr{Et = 0IEt - 1 = I}.
if E t ~ 1
(4)
Zt i = {hi(Xt) + Wi (Xt ) (11)
It is assumed that these probabilities are known a priori. 'wi otherWIse
However, if they are not known a very low value is assumed
(e.g., 0.01). where Wi (Xt ) is Gaussian distributed with zero mean and
The motion matrix, F, in (1) for a sampling interval of variance Ri(Xt ) and wi is Gaussian distributed with zero
T is given by mean and variance -f1 cat.
The complete measurement
recorded at time t is denoted as Zt = {Zt,i : i = 1 ... N},
1 T 0 0 0 and the set of all measurements up to time t is denoted
0 1 0 0 0 Zl:t = {Zk : k = 1 ... t}.
F= 0 0 1 T 0 (5)
0 0 0 1 0 3.3 Estimation Algorithm
0 0 0 0 1
The covariance matrix Q of Vt is given by The basic function of the particle filter is to approximate
the posterior density of the target state, given all measure-
T ql
3
T ql
2
0 0 0
2 ments, by a set of P points, X ~p) , called particles, and cor-
T~ql 0 0 0
2 Tql responding weights, w(p) [1]. That is
Q= 0 0 T 3 ql T 2 ql 0 (6)
-2- p
0 0
T~ql
2 Tql 0 p(XtIZ1 :t ) ~ L w~p)6(Xt - X~p»), (12)
0 0 0 0 Tq2 p=l

165

Authorized licensed use limited to: IEEE Xplore. Downloaded on October 30, 2008 at 00:59 from IEEE Xplore. Restrictions apply.
where 8(.) is the Dirac delta function. The particles and
their weights are updated recursively as new measurements
become available. At each time step the particle positions
are proposed from their previous position by sampling from
the target dynamics (1) and the weights are calculated using
the likelihood function arising from the measurement equa-
tion (11) assuming that the target is present
N (17)
w~p) ex II N {Zt, hi(Xt ), Ri(Xt )} .
i=1
where N { x, J-l, (J2} is the multivariate Gaussian function
with mean JL and variance (J2 evaluated at x.
Since the target motion model is linear in the target state
and the measurement model does not depend on the ve-
locities a technique known as Rao-Blackwellisation [1] or
marginalisation [10] can be used to update the velocities
of each particle. In this case each particle uses a standard
Kalman filter to update the velocities exactly, given the po-
sitions, rather than rely on Monte Carlo methods to explore
the velocity space. The method is one of a family of vari-
ance reduction techniques, which aim to reduce the variance
of the particle weights resulting in a more efficient filter.
At each time step the particle filter implemented here
forms two sets of particles. One set of particles carries
information about the target from the previous time step
(called the target particles), while the other set searches for a
new target in the data (called the birth particles). The prob-
ability of the target existing in the data, pte = Pr{ E t = 1},
can be calculated as a function of the weights of these two
sets of particles [11].
The resulting particle filter algorithm follows
1. Particle proposal (birth): A set of PB particles,
{X~B,p)}:~I' is generated assuming that there is no
target in the data. In this case we randomly place sam-
ples around the coverage region.
2. Particle proposal (target): A set of PT particles,
{X~T,p)}:~1 is proposed from the particles at the pre-
vious time, X ~~)I' sampling from (1) in combination
with the Rao-Blackwellised velocities.
3. Weight calculation: Once we have placed all the parti-
cles we need to compute their associated weights using
(13) and the measurement from the current time, Zt
II N { Zt, hi (XIS,p)), Ri(XIS,p))}
N
w~S,p) =
i=1
and then normalise
-(S,p) Ps
w(B,p) - ~ where WeB) - '""" iiJ(B,i)
t - weB) , t - L-t t
t i=1
where S can refer to either B or T.
Authorized licensed use limited to: IEEE Xplore. Downloaded on October 30, 2008 at 00:59 from IEEE Xplore. Restrictions apply.
4. Probability of existence: The probability of a target
existing in the data can be calculated using the sum of
the weights of the birth and target particles [11]
pe Q (16)
+ PdPte_ 1 + (1 - Pb)(1 - pte_I)
r-..J
t r-..J a
a = Wt(T) (1 - Pd)Pte_ 1 + Wt(B) Pb(1 - pte_I)'
(13)
5. Combining: The birth and target particles are then
combined into one set of PB + PT particles with
weights
w~B,p)* = Pb(1 - Pte_I)Wt(B)w~B,p) (18)
w~T,P)* = (1 - Pd)Pte_ 1 Wt(T)w~T,p), (19)
followed by another normalisation.
6. Resampling: The last step in the PF is the resampling
process. Resampling eliminates particles with weights
that are of low importance and multiplies those with
higher values. The resampling also reduces the set of
P B + PT particles down to PT , resulting in the set of
particles {X~P)}:~I all with uniform weights. We use
the systematic resampling procedure described in [1].
7. If Pe is above a predetermined threshold then a target
is declared present, where the particles resulting from
the previous step describe the target state. An estimate
of the target position can be calculated from the set of
particles by taking their mean
(20)
8. For the next time step, we collect a new set of readings
Zt+1 and go back to Step 1.
4 Simulation Results
As discussed earlier, one of the goals of this study is
to determine the effect of various parameters on the per-
formance of the tracking system. A secondary goal is to
provide recommendations on the choice of these network
design parameters to engineer a real WSN-based tracking
solution. The simulations were conducted using the NS2
(14)
discrete event simulator [5].
For our simulation studies, we assume that N sensors
are statically deployed in an area of size 120m x 120m.
We studied two deployment topologies: (i) Grid: where
(15) the nodes are placed at an equal distance from each other
resembling a perfect grid shape and (ii) Uniform Random
(UR): where the nodes are uniformly distributed over the
entire field of deployment. However, to avoid large areas
166
 120
102
90 85
75 ~ ~:.. • .. "0 . .~.. ... .,., ,;roo ~ oo. ."' ~
68
51
1
17
.~
o 17
15 30 45 00 75 90 105 12D
(a)
Figure 1. Grid and Uniform deployment topol-
ogy (49 nodes) with the actual and estimated
target trajectories.
without sensor coverage, we divided the area into a num-
ber of smaller cells of equal size (number of equal sized
cells depends on the total number of nodes that are being
deployed) and then randomly placed a node in each of the
resulting cells. Figure 1 illustrates the grid and uniform de-
ployment topologies with 49 nodes in the target area.
We assume that each node is equipped with an acoustic
sensor and that it samples the ambient sound at predefined
intervals. We refer to these intervals as time steps in the
rest of this report. In the simulator, we have modelled the
acoustic samples measured by the sensors as a simple dis-
tance based function, wherein the intensity of these readings
is an inverse function of the Euclidean distance between the
current position of the target and the location of the sen-
sors. The acoustic measurements recorded by the sensors
are, therefore, not actual readings but rather synthetically
generated based on the calibration results.
For simplicity, we limit this work to estimation of a sin-
gle target moving along a straight diagonal line from a point
at the bottom left comer of the topology to the right hand top
comer. The velocity of the target was set to 0.35m/sec and
each simulation was run for 450 seconds. The base station
is located at the centre of the topology (at location 60,60).
The complete simulation consists of two parts. The first part
is performed in NS2 where all the measurements during the
simulation run are recorded by each node and then these
measurements are forwarded to the centrally located base
station, over multi-hop, where they act as inputs to the PF.
The PF assumes the values ql = 0.002 and q2 = 10- 6 , in
reference to (6), and low values for the birth and death prob-
abilities, Pb = Pd = 0.01, in this case. Once all the data
is available, the base station runs the PF code (offline, in
MATLAB). The NS2 simulation setup parameters are listed
in Table 1.
Figure 1 shows the actual trajectory of the target and an
example of the estimated track predicted by the PF for both
Authorized licensed use limited to: IEEE Xplore. Downloaded on October 30, 2008 at 00:59 from IEEE Xplore. Restrictions apply.
Table 1. Simulation Parameters
J"
Parameter Value
~/ Radio Propagation Model Two Ray Ground
~/ Node transmission range 35m
~V Acoustic sensing range 25m
A' Speed of target
O.35m1sec
~r Region 120m x 120m
34 51 66 85 102 120
grid and DR deployment. As can be seen from this fig-
ure, the PF is very accurate in tracking the motion of the
(b) target. To quantitatively measure the performance of the
tracking system, we define a metric, accuracy ofestimation
as the average Euclidean distance between the actual and
estimated location of the target. This metric is computed
at each step of the execution of the PF. The second metric
used in our evaluation is the computation time required for
executing the PF algorithm. The total computation time is
T comp = T ns + Tpf where T ns is time (real-time) taken in
Ns2 to collect all the measurements at the base station for a
particular run of simulation and Tp f denotes the time taken
by MATLAB to run the PF at the base station.
For each run of the simulation, the measurements were
taken for each node at the sampling interval T. For each
time step when the PF detects the presence of the target,
the error is calculated in terms of the Euclidean distance
between the estimated location as indicated by the PF and
the actual target location. For each run of the simulation,
we calculated the statistical mean and standard deviation of
these errors. We repeated each of our simulations 100 times
and the results are shown as an average result across all the
runs including the standard deviation. A similar process is
repeated for the second metric, the computation time.
4.1 Calibration
Calibration of the nodes is an important aspect of a re-
alistic target tracking system. Referring to the mean of the
measurement function (9), repeated here, dropping the time
index for brevity
-
Zi -
2
ai + "Ii Id-~
i , (21)
in order to accurately track a target there are two parameters
which must be determined for each sensor i, that is the gain,
"Ii and the noise variance, a;. It is assumed that the other
values, ~ and c from (10) are known, or can be calculated
by a separate calibration mechanism, and are identical for
each mote. We assume that a white noise source of unit (ar-
bitrary) intensity is available, which can be placed at a unit
distance from the sensor to be calibrated. We also assume
that the resulting parameters are independent of the sensor
orientation, which mayor may not be valid, depending on
the sensor.
From (21), if there is no signal from the source (1 = 0 or
di ~ 00) then each measurement is solely from the noise,
167
 Grid Deployment Uniform Random Deployment CompLtation Time

- Grid Dep!oymen1
roo - + - UnIfonn Random D&ployment

500

300

16 25 36 49 64 81 100 16 25 36 49 64 81 100 16 25 36 49 64 81 100

Number of Nodes Number of Nodes Number of Nodes

(a) (b) (c)

Figure 2. Effect of number of deployed nodes on the estimation error and computation time

which encapsulates background noise and internal measure- these figures, for grid deployment the mean tracking error
ment noise. Thus, taking the mean of M o measurements drops from about 5.3 meters for 16 nodes to 2.47 meters
ziO,m) from sensor i gives an estimate of a; for 49 nodes. For DR deployment the corresponding de-
crease is from 5.37 meters to 2.45 meters. A high variation
Mo in mean and standard deviation is observed for DR deploy-
,,2 _ _1_ "" (O,m)
ai - M ~ zi . (22) ment. This can be attributed to the random nature of the
°m=l DR deployment strategy. For a fixed target trajectory, ran-
Similarly, if the source of unit acoustic power is placed at dom placement of nodes results in more variation in track-
exactly unit distance from the sensor, then 21 reduces to ing error across the simulation run as compared to the grid
deployment. The DR deployment performs better or worse
(1) 2 than the grid deployment depending on the node placement
Zi = ai + 1i (23)
with respect to the target trajectory.
and so if M 1 measurements are taken, then the sample mean
%age Packet Lost
gives an estimate of 1i
--II- Grid Deploymen1
.......... Uniform Random Deployment
20
(24)

For the simulations, gain and noise variations values are

randomly selected for each mote. The gain is chosen be-
tween 1000 and 8000 and the noise variance between 5 and
40. The remaining parameters in the measurement model
are chosen to be ~ = 2.2 (the path loss factor) and c = 1.5
(the measurement nonlinearity) as these values correspond 16 25 36 49 64 81 100
Num bar 01 No des
well to the measurements taken in the calibration procedure
before the experiments described in Section 6. Figure 3. Effect of the number of nodes on
packet loss.
4.2 Effect of the number of sensor nodes

To begin with, we focus our attention on evaluating the
effect of the number of deployed sensor nodes on the afore-
mentioned performance metrics. We increased the number
of sensor nodes from 16 to 100 in a 120m x 120m area.
Other important simulation parameters selected for this set
of simulations include fixing the number of particles to 5000
and the sampling interval equal to 1 sec. Figures 2 (a,b)
show the mean tracking error in the target's location for both
grid and uniform random topologies. As can be seen from
The mean tracking error for both grid and DR deploy-
ment fluctuates in a very small band for any further increase
in the number of the nodes beyond 49. However, if we
look at Figure 2(c), which shows the computation time of
the PF versus the number of deployed nodes, we can see
that the computation time of the PF grows almost linearly
with the number of deployed nodes for both grid and uni-
form random deployment. Packet loss also increase sharply
with increase in number of nodes beyond 49 as shown in

168

Authorized licensed use limited to: IEEE Xplore. Downloaded on October 30, 2008 at 00:59 from IEEE Xplore. Restrictions apply.
 Grid Deployment Uniform Random Deployment
Computation Tme

- - - Grid Deploymerrt
350 --+- Uniform Random Deployment

250

-
'- -
...........

- 200

150

100'-------1-'--'---'--'---'--'--'---'---'------'
1K 2K 3K 4K 5K 6K 7K 8K 9K 10K 1K 2K 3K 4K 51< 6K 7K 6K 9K 10K 1K 2K 3K 4K 51< 6K 7K 6K 9K 10K
Number 01 Particles Number of Particles Number of Particles

(a) (b) (c)

Figure 4. Effect of the number of generated particles on the estimation accuracy and computation
time.

Figure 3. This suggests, that for the given size of the de-
ployment field, target and sensor node characteristics (size,
velocity, and intensity of target, and sensitivity of sensors
etc.), and the selected networking and simulation parame-
ters, the optimal density for the deployment of sensor nodes
should be somewhere around 34 nodes per 100m2 , which
corresponds to 49 nodes for the target area under investiga-
tion. This node density gives a balance between the error
in estimation (mean tracking error of about 2.5m) and total
computation time (about 230 seconds).
4.3 Effect of the number of generated par-
ticles
Next, we examined the effect of the number of generated
particles on the performance of the tracking system. Re-
call that in our system, a number of particles are randomly
generated near the sensor nodes. Further, the number of par-
ticles are independent of the number of deployed nodes (see
Section 3) and only the Tpf component of the total compu-
tation time varies with change in number of particles.
Figure 4 shows the results of the simulation for differ-
ent number of generated particles for both grid and DR de-
ployments. The number of nodes are 49 and the sampling
interval is 1 sec. The error in estimation for grid deploy-
ment reduces from about 3.5 meters to 2.2 meters when the
number of particles are increased from 1000 to 6000. Af-
ter that the error begins to decrease at a much slower rate.
For DR deployment, the mean tracking error reduces from
3.49 meters with 1000 particles to 2.45 meters at 5000 par-
ticles. However, from Figure 4(c) we can see that the com-
putation time of the PF grows linearly with the number of
generated particles. This suggests that given the simulation
parameters, 5000-6000 particles is a good tradeoff between
detection accuracy and computation time.
4.4 Effect of the sampling interval
Figure 5, shows the effect of the variation in sampling
interval on the PF's behaviour. The number of nodes is set
as 49 with 5000 particles. Intuitively, the larger the sam-
pling interval, the longer the nodes will remain idle, hence
the less energy that they will consume. However, as can be
seen from this figure, the error in estimation of the target's
location begins to increase with the increase in the sampling
interval. Best tracking results are obtained with 0.5 to 1 sec
sampling interval for both grid and DR deployment. The
computation time, on the other hand, reduces from about
500 sec to 230 sec as the sampling interval is increased from
0.5 sec to 1 sec in Figure 5(c). This implies that given the
node density and the selected simulation parameters, 1 sec
sampling time is a sensible choice for this case.
5 Theoretical Analysis and Comparison
In this section we first derive the Cramer-Rao Bound for
the scenario used in our simulations. We next compare this
theoretical lower-bound with the results from our simula-
tions.
5.1 The Posterior Cramer-Rao Bound
(PCRB)
The Cramer-Rao Bound (CRB) is a theoretical construct,
which specifies a lower bound on the second-order estima-
tion error performance of any unbiased estimator [6]. We
now proceed to compute this theoretical bound for the sce-
nario used in our simulations (refer to Section 4 for details
on the simulation scenario).
The state evolution model and the measurement model
are as described in Section 3 (see also [12] for further de-

169

Authorized licensed use limited to: IEEE Xplore. Downloaded on October 30, 2008 at 00:59 from IEEE Xplore. Restrictions apply.
 Computation Tme
Grid Deployment (49 Nodes, 5000 Partides) UR Deploymeri (49 Nodes, 5000 Particles)
'0 ___ Grid Deployment
~ Uniform Random Deploymen1

'50

'00
50
OL--L..--L..--L..-'-----J'-----J~~~--------'--------' _, L---L..-L..-L..-L..-'------J'------Jl..--Jl..--Jl..--J--'

o , 2 3 4 5 6 7 8 9 10 " o , 2 3 4 5 6 7 8 9 '0 " , 2 3 4 5 6 7 8 9 '0

sampling Interval (in sec) sampling In1erval (in sec) sampling Interval (sec)

(a) (b) (c)

Figure 5. Effect of sampling interval on estimation accuracy and computation time.

tails). For simplicity we will limit our analysis to the sin-

gle target case, although the multi-target CRB directly fol- 0.7
-CRLB
-RMSError
lows [13]. As in the case of the simulation scenario, a WSN 0.6
composed of N = 49 acoustic sensor nodes is deployed in 0.5
a two-dimensional region of size 120m x 120m. E
~0.4
An unbiased estimate of the state vector Xt, based on t:
w
0.3
ZI:t and the known distribution of the initial target state,
p(Xo), is denoted by Xt, and its covariance matrix by Pt.
0.2

The lower bound of Pt, referred to as the posterior Cramer- 0.'

Rao bound is expressed as follows [6] 50 '00 '50 200 250 300 350 400 450
Time (sec)

Pt === E {(Xt - Xt)(Xt - Xtflzl:t,p(Xo)} ~ Jt- 1

Figure 6. Comparison of CRB and simulation
(25) results.
The inequality in (25) implies that the difference Pt - Jt- 1 is
a positive semi-definite matrix. Matrix Jt in Equation (25)
is referred to as the information matrix and its inverse is the
PCRB. once again evaluated at the true value of Xt. The terms in
For a deterministic nonlinear filtering problem specified (27) and (28) are as defined in Section 3.
by equations (1) and (11), the information matrix can be
computed recursively as follows [1] 5.2 Comparison of PCRB and Simulation
N
Results
J t+ 1 = [F-
1
]T J t F-
1
+ L H'!+I,i R t-Jl,i Ht +1 ,i, (26)
i-I For the PCRB calculations the initial state vector was
chosen as
where it is assumed that Q is zero. Here Ht,i is the Jacobian
of hi (Xt ) evaluated at the true value of X t X o = [15 0.25 5 0.25 0.7]T (29)

(Xi - Xt)~'iltd~ff-2 where the first and the third components are in meters, while
o the second and the fourth components are in meters/sec. For
Ht,i = (Yi - Yt)~'Yiltd~ff-2 (27) the purposes of this work the initial J 0 was approximated as
o a diagonal matrix with
-e
'Yi dt,i
J0 1 = Po = diag([O.l 0.05
0.01]2). 0.1 0.05
and R~; is given by (30)
The displayed error bounds are computed as follows:
(28)
(31)

170

Authorized licensed use limited to: IEEE Xplore. Downloaded on October 30, 2008 at 00:59 from IEEE Xplore. Restrictions apply.
 ....
)( ~

Figure 9. Particle Filter Convergence

Figure 7. Target and Experimental Setup
24 25
Acoustic Variance Vs Time (Route' ,2--'3, '4--23,24) @ @

18 2D
@ @,

~. 15
IS'

4000 ~o

5
@
3000

2000

'000

21 22
@ @

17
1
18
2 @ @

'-2
I"
~1
Figure 8. Acoustic Variance vs Time
(gI

7
~ 1 @

where J t- 1 [1, 1] and Jt- 1 [3, 3] are the diagonal elements of

the inverse of the information matrix corresponding to the
x and y coordinates respectively.
Next we compare the theoretical lower error bounds with
the root-mean square (RMS) errors of the simulated PF. The
number of particles used was 5000 and the PF was run 100
times for the bound comparison. Figure 6 illustrates that
the simulated PF follows the general trend of the theoretical
bound. We expect some mismatch between the filter per-
formance and the bound due to the fact that the bound is a
theoretically ideal quantity.
6 Initial Experiments
In order to test the performance of the PF based track-
ing system in a real world environment, we have conducted
some preliminary experiments. These experiments are also
used to validate the simulation results. In this section we
present our initial findings obtained from these experiments.
Figure 7 shows the setup of our testbed. We used a proto-
type testbed consisting of twenty five of the Xbow MicaZ
motes. These motes were programmed to perform high fre-
quency sampling to measure the acoustic signals (at 5kHz)
generated by the target (a remote controlled toy car). These
MicaZ motes were deployed in a grid of size 2.5 meters x
2.5 meters (Figure 7) with 50cm grid spacing. A mote at-
tached with a laptop was used as the data sink. The motes
measure the acoustic signals and send a summary of the
acoustic samples to the laptop (single hop communication),
which executes the PF. The approximate speed of the target
Figure 10. Tracking Results
was 0.03rn/sec.
Time synchronisation among the motes was achieved by
a Beeper mote broadcasting a beacon message every 0.5
seconds (sampling interval). On receiving a beacon, each
mote starts sampling to collect 1000 discrete samples (tak-
ing 0.2s at 5kHz). To reduce the storage requirements at the
motes, radio communications and the size of the data pack-
ets' motes only maintain summary statistics (sum and sum
of squares) instead of the raw values. These summary statis-
tics were then transferred during the next 0.3 seconds over
the wireless link to the base station. Note that the sample
variance can be easily calculated at the base station as we
know the total number of samples, sum, and sum of squares
of the samples, as shown
(32)
To ensure that all data packets are received at the
base station, we incorporated an application level TDMA
scheme to allocate discrete contention-free time slots to
each mote for data transfer to the base station. The base sta-
tion executes the PF with 5000 particles. Prior to running
the experiments, the environmental path loss factor and the

171

Authorized licensed use limited to: IEEE Xplore. Downloaded on October 30, 2008 at 00:59 from IEEE Xplore. Restrictions apply.
standard deviation of the background noise are estimated
by calibrating each mote carefully in the same experimental
environment, as described in Section 4.1.
Figure 8 shows the acoustic variance with time calcu-
lated at the base station for some of the deployed sensors.
Figure 9 illustrates the two states of the PF, where the ini-
tial set of particles are shown searching for a new target
on the left, and particles, after a number of measurements
from the target, are shown on the right. Some of the track-
ing results are shown in Figure 10 where actual paths are
shown in black arrows. The figures only show the quali-
tative performance of the WSN based tracking system. We
have not used the quantitative metrics described in the simu-
lation (Section 4) as fine-grained ground truth about the po-
sition of the target, both in time and space, was not available
for error computation. The qualitative results demonstrate
that the PF based estimator performs very well in tracking
different trajectories of the target using real data collected
by off-the shelf WSN devices.
7 Conclusion and Future Work
In this paper we present a detailed simulation-based eval-
uation of the performance of a WSN tracking system, which
employs the PF algorithm. We observed the effect of vari-
ous network design parameters such as number of deployed
nodes, number of particles, and sampling interval on the
estimation accuracy and computation time of the tracking
system. We found that, for the given Ns2 simulation param-
eters, the optimal deployment density is around 34 nodes
per 100m2 , optimal number of generated particles is about
5000, and optimal sampling interval is 1 sec. These recom-
mendations on the choice of network design parameters can
be utilized to engineer a real WSN-based tracking solution.
We also derived the theoretical lower bound, the Cramer-
Rao Bound (CRB) and demonstrated that the results from
our simulations are comparable to the bound. We presented
initial results on the performance of a prototype WSN track-
ing system with 25 sensor nodes. The PF based tracking
system performs well in tracking targets performing differ-
ent manoeuvres in the target area.
We are currently conducting further experiments to test
our system in real environments. Our future goals are to en-
able real-time distributed tracking of multiple targets and to
incorporate a localisation algorithm in the tracking system.
Acknowledgement
This work is funded by the Defence Science and Tech-
nology Organisation (DSTO), Australia.
References
[1] B. Ristic, S. Arulampalam, and N. Gordon, Beyond
the Kalman Filter: Particle Filters for Tracking Appli-
172
Authorized licensed use limited to: IEEE Xplore. Downloaded on October 30, 2008 at 00:59 from IEEE Xplore. Restrictions apply.
cations. Artech House, 2004.
[2] L. Gu, D. Jia, P. Vicaire, T. Yan, L. Luo, A. Tirumala,
Q. Cao, T. He, J. A. Stankovic, T. Abdelzaher, and
B. H. Krogh, "Lightweight detection and classifica-
tion for wireless sensor networks in realistic environ-
ments," in Proceedings of ACM SenSys 2005, 2005,
pp.205-217.
[3] M. F. Duarte and Y. H. Hu, "Vehicle classification in
distributed sensor networks," J. Parallel Distrib. Com-
put, vol. 64, no. 7,pp.826-838,2004.
[4] Q. Wang, W. Chen, R. Zheng, K. Lee, and L. Sha,
"Acoustic target tracking using tiny wireless sensor
devices," in 2nd Inti. Conf. on Information Processing
in Sensor Networks (IPSN03), 2003.
[5] NS2, "Network simulator 2,"
http://www.isi.com/nsnam/ns/.
[6] H. L. V. Trees, Detection, Estimation and Modulation
Theory (Part I). John Wiley and Sons, 1968.
[7] A. Ledeczi, A. Nadas, P. Volgyesi, G. Balogh,
B. Kusy, J. Sallai, G. Pap, S. Dora, K. Molnar,
M. Maroti, and G. Simon, "Countersniper system for
urban warfare," ACM Transaction Sensor Networks,
vol. l,no.2,pp. 153-177,2005.
[8] T. He, S. Krishnamurthy, J. A. Stankovic, T. F. Ab-
delzaher, L. Luo, R. Stoleru, T. Yan, L. Gu, J. Hui,
and B. Krogh, "An energy-efficient surveillance sys-
tem using wireless sensor networks," in Proceedings
ofACM Mobisys, 2004.
[9] M. 1. Coates and G. lng, "Sensor network particle
filters: motes as particles," in Proceedings of IEEE
Workshop on Statistical Signal Processing, Bordeaux,
France, 2005.
[10] T. Schon, F. Gustafsson, and P.-J. Nordlund,
"Marginalized particle filters for mixed lin-
ear/nonlinear state-space models," IEEE Trans.
Signal Processing, 2005.
[11] M. G. Rutten, N. J. Gordon, and S. Maskell, "Recur-
sive track-before-detect with target amplitude fluctua-
tions," lEE Proc.-Radar Sonar Navig., vol. 152, no. 5,
pp. 345-352, Oct. 2005.
[12] B. Ristic and S. Arulampalam, "Integrated detection
and tracking of multiple objects with a sensor net-
work," in Preprint.
[13] B. Ristic and M. Morelande, "Cramer-rao bound for
multiple target tracking using intencity measurments,"
in Conference on Information, Decision and Control
(IDC 2007), Adelaide, Australia, Feb. 2007.