Professional Documents
Culture Documents
A Scalable Multi-Target Tracking Algorithm For WSNs
A Scalable Multi-Target Tracking Algorithm For WSNs
A Scalable Multi-Target Tracking Algorithm For WSNs
Research Article
A Scalable Multi-Target Tracking Algorithm for
Wireless Sensor Networks
Songhwai Oh
School of Electrical Engineering and Computer Science, ASRI, Seoul National University, Seoul 151-744, Republic of Korea
Copyright © 2012 Songhwai Oh. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Multi-target tracking is a representative real-time application of sensor networks as it exhibits different aspects of sensor networks
such as event detection, sensor information fusion, multihop communication, sensor management, and real-time decision making.
The task of tracking multiple objects in a wireless sensor network is challenging due to constraints on a sensor node such as
short communication and sensing ranges, a limited amount of memory, and limited computational power. In addition, since a
sensor network surveillance system needs to operate autonomously without human operators, it requires an autonomous real-
time tracking algorithm which can track an unknown number of targets. In this paper, we develop a scalable real-time multi-target
tracking algorithm that is autonomous and robust against transmission failures, communication delays, and sensor localization
error. The algorithm is based on a rigorous probabilistic model and an approximation scheme for the optimal Bayesian filter. In
particular, an extensive simulation study shows that there is no performance loss up to an average localization error of 0.7 times the
separation between sensors and the algorithm tolerates up to 50% lost-to-total packet ratio and 90% delayed-to-total packet ratio.
The proposed algorithm has been successfully applied to real-time multi-target tracking problems using wireless sensor networks.
is closest, thereby selecting a single association. Although with previously received observations to improve the accu-
effective under benign conditions, the NNF gives order- racy of estimates. Furthermore, MCMCDA requires less
dependent results and breaks down under more difficult memory as it maintains only the current hypothesis and the
circumstances. Some algorithms are designed for multi- hypothesis with the highest posterior. It does not require the
target tracking [22–24] where the complexity of the data enumeration of all or some of hypotheses as in [7, 28]. In
association problem inherent to multi-target tracking is this paper, we extend the MCMCDA algorithm to sensor
avoided by classification [22, 23] or heuristics [22, 24]. The networks in a hierarchical manner so that the algorithm
associations between measurements and targets are not gen- becomes scalable. To our knowledge, the algorithm presented
erally known in multi-target tracking. The data association in this paper is the first general real-time scalable multi-
problem is to determine which measurements were gener- target-tracking algorithm for sensor networks which can
ated by which targets. When tracking targets of a similar type systematically track an unknown number of targets in the
or when reliable classification information is not available, presence of false alarms and missing observations and is
the classification-based tracking algorithm behaves as the robust against transmission failures, communication delays,
NNF. Considering the fact that the complexity of the data and sensor localization error.
association problem is NP-hard [25, 26], a heuristic approach
We consider a simple shortest-path routing scheme on a
breaks down under difficult circumstances. Furthermore, the
sensor network. The transmission failures and communica-
measurement inconsistencies common in sensor networks,
tion delays of the network are characterized probabilistically.
such as false alarms and missing measurements due to
We assume the availability of a small number of special
missing detection or packet loss, are not fully addressed in
many algorithms. On the contrary, the multi-target tracking nodes, supernodes, that are more capable than regular nodes
algorithm developed in this paper is based on a rigorous in terms of computational power and communication range.
probabilistic model and based on a true approximation Examples of a supernode include high-bandwidth sensor
scheme for the optimal Bayesian filter. nodes such as iMote and BTnode [29], gateway nodes such
We can categorize a target tracking algorithm for sensor as Stargate, Intrinsyc Cerfcube, and PC104 [29], and the
networks by its computational structure: centralized [11, 14, Tier-2 nodes used in the experiment of [27]. Each node is
15], hierarchical [16, 17], or distributed [10, 12, 13, 18–24]. assigned to its nearest supernode and nodes are grouped
Since each sensor can only sense a small region around it, and by supernodes. We call the group of sensor nodes formed
its measurements are noisy and inconsistent, measurements around a supernode a “tracking group”. When a node
only from a single sensor and its neighboring sensors are not detects a possible target, it communicates with its neighbors
sufficient to initiate, maintain, disambiguate, and terminate and observations from the neighboring sensors are fused
tracks of multiple targets. It requires measurements from and sent to its supernode. Each supernode receives the
distant sensors to initiate, maintain, disambiguate, and ter- fused observations from its tracking group and executes the
minate tracks of multiple targets. But considering the com- tracking algorithm. Each supernode communicates with
munication load and delay when exchanging measurements neighboring supernodes when a target moves away from
between distant sensors, a completely distributed approach its range. Lastly, the tracks estimated by supernodes are
is not feasible for real-time applications. On the other hand, combined hierarchically. Although a specific sensor network
a completely centralized approach is not robust and scalable. model is used for the performance evaluation, the algorithm
Hence, we consider a hierarchical architecture to minimize is applicable for different routing algorithms and sensor
the communication load and delay while being robust models, for example, distributed air traffic control [30].
and scalable. In fact, our algorithm has been successfully As mentioned earlier, the algorithm presented in this
demonstrated using a large-scale outdoor sensor network paper is used in [27]. Although the overview of the algorithm
deployment and used to solve a real-time control problem is described in [27], the focus of [27] is to describe a new
[27]. control system architecture using sensor networks and an
Markov chain Monte Carlo data association (MCMCDA) outdoor experiment. In this paper, we describe the algorithm
presented in [9] can track an unknown number of targets in in detail and study the characteristics of the algorithm under
real-time and is an approximation to the optimal Bayesian different conditions in simulation. The paper also includes
filter. It has been shown that MCMCDA is computationally experimental results from an indoor experiment with 45
efficient compared to the multiple hypothesis tracker (MHT) sensors and ourdoor experiments described in [27]. While
[7] and outperforms MHT with heuristics (i.e., pruning, a wireless sensor network generates noisy, inconsistent, and
gating, clustering, N-scan-back logic and k-best hypotheses) bursty measurements, the algorithm described in this paper
under extreme conditions, such as a large number of targets successfully estimated the number of targets and tracks of
in a dense environment, low detection probabilities, and targets in real-time.
high false alarm rates [9]. MCMCDA is suitable for sensor The remainder of this paper is structured as follows. The
networks since it can autonomously initiate and terminate sensor network model assumed in this paper is described in
tracks. Since transmission failure is another form of miss- Section 2. In Section 3, the multi-target tracking problem,
ing observation, MCMCDA is robust against transmission the MCMCDA algorithm, and the hierarchical extension of
failures. MCMCDA performs data association based on both MCMCDA are described. The simulation and experiment
current and past observations, so delayed observations, that results are given in Sections 4 and 5. Frequently appearing
is, out-of-sequence measurements, can be easily combined symbols are listed in Abbreviations section.
International Journal of Distributed Sensor Networks 3
between the number of delayed packets, and the number of the state variable can be extended to include target type
total packets as a function of the communication delay rate information as done in [32].
pde . The main objective of the multi-target tracking problem
When the network is heavily loaded, the independence is to estimate K, {tik , t kf } and {xk (t) : tik ≤ t ≤ t kf }, for k =
assumptions on transmission failure and communication 1, . . . , K, from noisy observations.
delay may not hold. However, the model is realistic under the Let Y (t) = { y j (t) : j = 1, . . . , n(t)} be all measurements
moderate conditions, and we have chosen it for its simplicity. at time t and Y = {Y (t) : 1 ≤ t ≤ T } be all measurements
from t = 1 to t = T. Let Ω be a collection of partitions of Y
3. Multi-Target Tracking such that, for ω ∈ Ω, ω = {τ0 , τ1 , . . . , τK }, where τ0 is a set of
false alarms and τk is a set of measurements from target k for
3.1. Problem Formulation. Let T ∈ Z+ be the duration of k = 1, . . . , K. Note that ω is also known as a joint association
surveillance. Let K be the number of targets that appear in event in literature. More formally, ω is defined as follows:
the surveillance region R during the surveillance period.
Each target k moves in R for some duration [tik , t kf ] ⊂ (1) ω = {τ0 , τ1 , . . . , τK };
K
[1, T]. Notice that the exact values of K and {tik , t kf } are (2) k=0 τk = Y and τi ∩ τ j = ∅ for i =
/ j;
unknown. Each target arises at a random position in R at tik , (3) τ0 is a set of false alarms;
moves independently around R until t kf , and disappears. At (4) |τk ∩ Y (t)| ≤ 1 for k = 1, . . . , K and t = 1, . . . , T;
each time, an existing target persists with probability 1 − pz (5) |τk | ≥ 2 for k = 1, . . . , K.
and disappears with probability pz . The number of targets
arising at each time over R has a Poisson distribution with a An example of a partition is shown in Figure 2. Here, K is the
parameter λb V where λb is the birth rate of new targets per number of tracks for the given partition ω ∈ Ω. We call τk a
unit time, per unit volume, and V is the volume of R. The track when there is no confusion although the actual track
initial position of a new target is uniformly distributed over is the set of estimated states from the observations τk . This
R. is because we assume there is a deterministic function that
Let F k : Rnx → Rnx be the discrete-time dynamics of the returns a set of estimated states given a set of observations. A
target k, where nx is the dimension of the state variable, and track is assumed to contain at least two observations since we
let xk (t) ∈ Rnx be the state of the target k at time t for t = cannot distinguish a track with a single observation from a
tik , . . . , t kf . The target k moves according to false alarm, assuming λ f > 0. For special cases in which pd =
1 or λ f = 0, the definition of Ω can be adjusted accordingly.
Let ne (t − 1) be the number of targets at time t − 1, nz (t)
xk (t + 1) = F k xk (t) + wk (t), for t = tik , . . . , t kf − 1,
be the number of targets terminated at time t and nc (t) =
(5) ne (t − 1) − nz (t) be the number of targets from time t − 1
that have not terminated at time t. Let nb (t) be the number
where wk (t) ∈ Rnx are white-noise processes. The white- of new targets at time t, nd (t) be the number of actual target
noise process is included to model nonrectilinear motions detections at time t, and nu (t) = nc (t) + nb (t) − nd (t) be the
of targets. When a target is present, a noisy observation number of undetected targets. Finally, let n f (t) = n(t) − nd (t)
(or measurement) of the state of the target is measured be the number of false alarms. Using the Bayes rule, it can be
with a detection probability pd . Notice that, with probability shown that the posterior of ω is:
1 − pd , the target is not detected and we call this a missing
observation. There are also false alarms, and the number P(ω | Y ) ∝ P(ω) · P(Y | ω)
of false alarms has a Poisson distribution with a parameter
λ f V , where λ f is the false alarm rate per unit time, per
T
nc (t) nu(t)
unit volume. Let n(t) be the number of observations at ∝ pznz (t) 1 − pz pdnd (t) 1 − pd (7)
t =1
time t, including both noisy observations and false alarms.
Let y j (t) ∈ Rn y be the jth observation at time t for j =
n f (t)
nb (t)
1, . . . , n(t), where n y is the dimension of each observation × (λb V ) λf V · P(Y | ω),
vector. Each target generates a unique observation at each
where P(Y | ω) is the likelihood of observations Y given
sampling time if it is detected. Let H j : Rnx → Rn y be the
ω, which can be computed based on the chosen dynamic
observation model. Then the observations are generated as
and measurement models. For example, the computation of
follows:
⎧
P(Y | ω) for the linear dynamic and measurement models
⎪
⎨H j xk (t) + v j (t), if y j (t) is from xk (t), can be found in [9]. Our formulation of (7) is similar
y j (t) = ⎪ (6) to MHT [28], and the derivation of (7) can be found in
⎩u f (t), otherwise, [28]. The parameters pz , pd , λb , and λ f have been widely
used in many multi-target-tracking applications [6, 28].
where v j (t) ∈ Rn y are white noise processes and u f (t) ∼ Our experimental and simulation experiences show that
Unif(R) is a random process for false alarms. We assume our tracking algorithm is not sensitive to changes in these
that the targets are indistinguishable in this paper, but if parameters in most cases. In fact, we used the same set of
observations include target type or attribute information, parameters for all our experiments.
International Journal of Distributed Sensor Networks 5
4. Simulation Results
Birth
For simulations below, we consider the surveillance over a
Death rectangular region on a plane, R = [0, 100]2 . The state vector
is x = [x1 , x2 , ẋ1 , ẋ2 ]T , where (x1 , x2 ) is a position in R along
(a) (b)
the usual x and y axes and (ẋ1 , ẋ2 ) is a velocity vector. The
Split
linear dynamic and measurement models are used
x(t + δ) = Aδ x(t) + Gδ w(t),
Merge (9)
(d)
y(t) = Cx(t) + v(t),
(c)
where δ is a sampling interval, w(t) and v(t) are white Gaus-
Ext. sian noises with zero mean and covariance Q = diag(0.152 ,
0.152 ) and R (set according to Figure 1), respectively, and
Red.
⎡ ⎤
⎡ ⎤
δ2
(e) (f) ⎢2 0⎥
1 0 δ 0 ⎢ ⎥
⎢0 1 0 δ ⎥ ⎢ δ2 ⎥
Update ⎢ ⎥ ⎢ ⎥
Aδ = ⎢ ⎥, Gδ = ⎢
⎢
0 ⎥, (10)
⎣0 0 1 0 ⎦
⎢ 2⎥ ⎥
0 0 0 1 ⎣δ 0⎦
0 δ
(g) (h)
1 0 0 0
Switch C= . (11)
0 1 0 0
0.4 0.18
0.35 0.16
Estimation error (number of targets)
0.12
0.25
0.1
0.2
0.08
0.15
0.06
0.1
0.04
0.05 0.02
0 0
1 2 3 4 5 0 0.5 1 1.5 2
Sensing range (Rs ), 10 targets Sensor localization error (σ), 10 targets
1.8 3
1.6 2.5
Estimation error (position)
1.4 2
1.2 1.5
1 1
0.8 0.5
0.6 0
0 0.5 1 1.5 2
0.4 Sensor localization error (σ), 10 targets
(b)
0.2
1 2 3 4 5
Sensing range (Rs ), 10 targets Figure 5: (a) The estimation error in the number of targets K as
a function of the localization error (σ). (b) The estimation error in
Slow speed position X as a function of the localization error (σ).
Medium speed
Fast speed
(b)
from target positions. Figure 4 shows the estimation errors
Figure 4: (a) The estimation error in the number of targets K as K and X when 10 targets appear and disappear at random
a function of sensing range Rs . (b) The estimation error in position times and T = 50. For slow-speed vehicles, |ẋ1 |, |ẋ2 | ∈ [0, 1];
X as a function of sensing range Rs . Unity corresponds to the sepa-
|ẋ1 |, |ẋ2 | ∈ [1, 2] for medium-speed vehicles; |ẋ1 |, |ẋ2 | ∈
ration between sensors.
[2, 5] for high-speed vehicles. For each vehicle type, five
different scenarios are used. When Rs = 0.5, the sensors
do not completely cover the surveillance region R and do
an example of surveillance using wireless sensor networks not detect targets at all times, hence the estimation error is
and demonstrate how the hierarchical MCMCDA algorithm higher. But as we increase Rs beyond 1.0, estimation errors
works. increase, since there are more collisions among observations
of different targets. The estimation errors are low for high-
4.1. Sensing Range. When localizing a single target, we can speed vehicles since it is easier to disambiguate crossing
minimize the localization error by allowing more sensors to targets. From Figure 4, we find that Rs = 1.5 is a good range
collaborate, which is equivalent to increasing Rs as shown for all types of vehicles and it is used in simulations below.
in Figure 1. But when there is more than one target, this is Notice that when Rs = 1.5, for each sensor si , there are 28
no longer true, since observations from different targets can neighboring sensors which are at most 2Rs away from si . We
collide, giving missing observations and observations away can also interpret this result in terms of sensor density for
8 International Journal of Distributed Sensor Networks
1 0.8
0.7
0.8
0.6
0.6 0.5
0.4
0.4
0.3
0.2
0.2
0.1
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Transmission failure rate, 10 targets Transmission failure rate, 10 targets
(a) (b)
0.7
0.6
Estimation error (position)
0.5
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
Transmission failure rate, 10 targets
(c)
Figure 6: (a) Ratio between the number of lost packets and the number of total packets. (b) The estimation error in the number of targets
K as a function of pte . (c) The estimation error in position X as a function of pte .
a fixed value of Rs . Hence, once the surveillance region is error is 0.707 times the separation between sensors. Notice
fully covered by sensors, a further increase in density does that K is always under 0.18, so the algorithm finds most
not improve the estimation error. tracks for all σ. But X gets larger at high σ, since the target
position estimates are based on incorrect node positions.
Considering the fact that X is computed from the norm of a
4.2. Sensor Localization Error. The localization of sensor vector in R2 , X is mostly due to the sensor localization error.
nodes in an adhoc wireless sensor network, without expen-
sive hardware such as the global positioning system (GPS),
is a challenging problem [34]. Hence, an algorithm which 4.3. Transmission Failures. To assess the effects of transmis-
utilizes sensor positions needs to be robust against the sensor sion failures alone, we assume that there are no delayed
localization error. Suppose that the true position of sensor observations, no false alarms, and no missing detections.
node i is s∗i and si = s∗i + wil , where wil are Gaussian A single supernode is placed at the center. As mentioned
noises with zero mean and covariance Σ = diag(σ 2 , σ 2 ). earlier, transmission failures are missing observations and
Figure 5 shows the estimation errors from tracking 10 targets Figure 6(a) shows the ratio between the number of lost
as functions of the sensor localization error σ. It shows that packets and the number of total packets as a function of
the algorithm is robust against the sensor localization error the transmission failure rate pte . As pte increases, we lose
and, for σ ≤ 0.5, the algorithm performs as if there is no more packets and, at pte ≈ 0.9, we lose all packets. Figure 6
sensor localization error. This is remarkable since σ ≤ 0.5 shows the estimation errors and the algorithm performs well
corresponds to the case in which the average localization for pte ≤ 0.4. The estimation error X is low at high pte
International Journal of Distributed Sensor Networks 9
1 0.7
0.8
0.5
0.6
0.4
0.3
0.4
0.2
0.2
0.1
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Communication delay rate, 10 targets Communication delay rate, 10 targets
Delayed
Delayed but received
(a) (b)
0.34
0.32
Estimation error (position)
0.3
0.28
0.26
0.24
0.22
0 0.2 0.4 0.6 0.8 1
Communication delay rate, 10 targets
(c)
Figure 7: (a) Ratio between the number of delayed packets and the number of total packets. (b) The estimation error in the number of
targets K as a function of pde . (c) The estimation error in position X as a function of pde .
since most of packets are lost at high pte , making the data Figure 7(a) shows the ratio between the number of delayed
association problem easier. Notice that when pte = 0.4 more packets and the number of total packets as a function of
than 50% of packets are lost (see Figure 6(a)). It shows that the communication delay rate pde . As pde increases to 1, all
our algorithm is very robust against transmission failures packets are delayed. Since ws = 10, we do not receive all
and tolerates up to 50% lost-to-total packet ratio. Hence, for the delayed packets and the ratio between the number of
given pte , we can find the maximum radius kmax of a tracking delayed packets that are eventually received and the number
group, measured by the number of hops from a supernode, of packets is shown in a dotted line in Figure 7(a). The
such that pts (kmax ; pte ) ≥ 0.5, sustaining the performance estimation errors are shown in Figure 7. It shows a good
of the algorithm. For example, if pte = 0.1, kmax = 38, performance for pde ≤ 0.6. At pde = 0.6, 90% of packets are
that is, the most distant node can be 38 hops away from a delayed and 72% of packets have delays less than ws . Hence,
supernode. But one must consider the communication delay the system tolerates up to 90% delayed-to-total packet ratio.
which increases with the number of hops.
4.5. An Example of Surveillance with Sensor Networks. In
4.4. Communication Delays. To assess the effects of commu- this section, we give an example of surveillance with sensor
nication delays alone, we assume that there are no transmis- networks. The surveillance region R is divided into four
sion failures, no false alarms, and no missing observations. quadrants and sensors in each quadrant form a tracking
10 International Journal of Distributed Sensor Networks
100 100
47
90 90
27 23
80 9 4 80
70 70
82
46 100
60 60
50 70 50
y
y
100
53 51 60
40 40
56
30 30
3
20 100 20
10 74 10
90
0 1 13 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
x x
(a) (b)
100 100
90 90
80 80
70 70
60 60
50 50
40 40
30 30
20 20
10 10
0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
(c) (d)
Figure 8: (a) A scenario used in Section 4.5 (numbers are target appearance and disappearance times, initial positions are marked by
circles). (b) Accumulated observations received by supernodes with delayed observations circled. (c) Tracks estimated locally by supernodes
superimposed. (d) Tracks estimated by the track-level data association step.
group, where a supernode is placed at the center of each that the track-level data association step corrects mistakes
quadrant. The scenario is shown in Figure 8(a). We used made by supernodes due to missing observations at the
pte = 0.3, pde = 0.3, and η = 2. The surveillance duration tracking group boundaries. The ability to correct mistakes
is increased to T = 100. Figure 8(b) shows the observations made by a supernode is another strength of our algorithm.
received by supernodes. There were a total of 1,174 observa- The algorithm is written in C++ and MATLAB and run on
tions and 603 observations were false alarms. A total of 319 a PC with a 2.6-GHz Intel Pentium 4 processor. It takes less
packets out of 1,174 packets were lost due to transmission than 0.06 seconds per supernode, per simulation time step.
failures and 449 packets out of 855 received packets were
delayed. The tracks estimated by the algorithm are shown 5. Experiments
in Figure 8(c). Figure 8(c) shows the tracks estimated locally
by supernodes while Figure 8(d) shows the tracks estimated 5.1. Indoor Experiments. As a proof of concept, we imple-
by the track-level data association step. Figure 8(d) shows mented a scaled down tracking scenario on a 9-by-5 sensor
International Journal of Distributed Sensor Networks 11
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
0 1 2 3 4 5 6 7 8 9 10 [5pt] 0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
(d) (t = 24.2 sec) (e) (t = 25.3 sec) (f) (t = 29.7 sec)
Figure 10: (a)–(c) Snapshots from the experiment of crossing tracks, proceeding left to right. We are looking at the grid of nodes from the
bottom right corner. (d)–(f) Corresponding tracks found by the MCMCDA algorithm at each time step. Reporting sensor nodes are circled
in red, postulated tracks are represented by dashed lines, while established tracks are represented by red solid lines. Note that the algorithm
makes a bad association at t = 25.3 sec, but corrects for it by t = 29.7 sec.
6 6
5 5
(21)
4 4
(22)
3 3
(22) (23) (23) (25) (26)
2 2 (25) (29)
1 1 (26)
0 0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
(a) (b)
Figure 11: (a) Trajectories of targets from the experiment. (b) Estimated tracks from the algorithm with observations (circles).
with its neighboring Tier-2 nodes when a target moves away the estimated time to capture the last evader is minimized.
from the region monitored by its tracking group. Lastly, Once the assignments are determined, the path planner
the tracks estimated by the Tier-2 nodes are combined module computes a trajectory for each pursuer to capture its
hierarchically by the MTF module at the base station. assigned evader in the least amount of time without colliding
The estimates computed by the MTF module are then into other pursuers. Then, the base station transmits each
used by the MAC module to estimate the expected capture trajectory to the path-following controller of the correspond-
times of all pursuer-evader pairs. Based on these estimates, ing pursuer. The path-following controller modifies the
the MAC module assigns one pursuer to one evader by pursuer’s trajectory on the fly to avoid any obstacles sensed
solving the bottleneck assignment problem [42] such that by the pursuer’s on-board sensors. The MTT and MTF
International Journal of Distributed Sensor Networks 13
(a) (b)
Figure 12: Hardware for the sensor nodes. (a) Triosensor node on a tripod. On top is the microphone, buzzer, solar panel, and user and reset
buttons. On the sides are the windows for the passive infrared sensors. (b) A live picture from the two-target PEG experiment. The targets
are circled.
Pursuer N
Input to the field around time 10 and 80, respectively. During the
Pursuers
Pursuer 1 Pursuer 2
pursuer thrusters
Path follower Path follower ...
Path follower
experiment, the algorithm correctly rejected false alarms
Collision-free desired
and compensated for missing detections. There were many
trajectories false alarms during the span of the experiments. Though
Path planner not shown in the figures, the algorithm dynamically cor-
Base station
Assignment of
pursuers to evaders
rected previous track hypotheses as it received more sensor
Multi agent
coordination (MAC) readings. A live picture of this experiment is shown in
Number of evaders,
position, velocities, and Figure 12(b). Figures 14(b) and 14(c) show the multi-target-
Multi track estimation error bounds
tracking results with two persons walking through the field.
fusion (MTF)
Regional estimates: Note that the MTT module is able to correctly disambiguate
number of evaders,
the presence of two targets (right panel of Figure 14(c)) using
Sensor network
Multi target ... Multi target past measurements, despite the fact that the MSF module
tracking (MTT) tracking (MTT)
Local estimation of reports the detection of a single target (upper left panel of
evaders positions from
sensor network Figure 14(b)).
Multi-sensor fusion (MSF)
Tier-1
Sensor
measurements 6. Conclusions
Figure 13: LochNess: a hierarchical real-time control system archi- In this paper, a scalable hierarchical multi-target-tracking
tecture using wireless sensor networks for multi-target tracking and algorithm for wireless sensor networks is presented. The
multiagent coordination. algorithm is based on the efficient MCMCDA algorithm,
and it is suitable for autonomous real-time surveillance in
sensor networks. This new multi-target-tracking algorithm
modules are based on the proposed hierarchical multi-target- can initiate and terminate tracks and requires a small amount
tracking algorithm. For more detail about the description of memory. The task of tracking is done hierarchically for
about the system used in the experiments, see [27]. real-time operation by forming a tracking group around a
We demonstrated the system on one, two, and three supernode and later combining tracks from different supern-
human targets, with targets entering the field at different odes. In order to reduce the communication overhead,
times. In all three experiments, the tracking algorithm cor- observations are first fused locally and then transmitted to
rectly estimated the number of targets and produced cor- its supernode. The algorithm is robust against sensor local-
rect tracks. Furthermore, the algorithm correctly disam- ization error and there is no performance loss up to the
biguated crossing targets in the two-target and three-target average localization error of 0.7 times the separation between
experiments without classification labels on the targets. sensors. The algorithm is also robust against transmission
Figure 14(a) shows an example with three humans walking failures and communication delays. In particular, the algo-
through the field. The three humans entered and exited rithm tolerates up to 50% lost-to-total packet ratio and 90%
14 International Journal of Distributed Sensor Networks
Detection Estimated tracks and pursuer-to-evader assignment Detection Estimated tracks and pursuer-to-evader assignment Estimated tracks and pursuer-to-evader assignment
90 −10 90 −10 90 Detection −10
80 80 80
70 −20 70 −20 70 −20
60 60 60
50 50 50
−30 −30 −30
40 40 40
30 30 30
20 −40 −40 20 −40
20
10 10 10
0 −50 0 −50 0 −50
0 20 40 60 0 20 40 60 0 20 40 60
Fusion Fusion Fusion
−60 −60 −60
80 80 80
−70 −70 −70
60 60 60
40
−80 −80 40
−80
40
Figure 14: (a) Estimated tracks of targets at time 70 from the experiment with three people walking in the field. (upper left) Detection
panel. Sensors are marked by small dots and detections are shown in large disks. (lower left) Fusion panel shows the fused likelihood. (right)
Estimated tracks. (b) Estimated tracks of targets from the experiment with two people walking in the field before crossing. (c) Estimated
tracks of targets from the experiment with two people walking in the field after crossing.
[11] A. Arora, P. Dutta, S. Bapat et al., “A line in the sand: a [26] A. Poore, “Multidimensional assignment and multitarget
wireless sensor network for target detection, classification, and tracking,” in Partitioning Data Sets, I. J. Cox, P. Hansen, and
tracking,” Computer Networks, vol. 46, no. 5, pp. 605–634, B. Julesz, Eds., pp. 169–196, American Mathematical Society,
2004. Providence, RI, USA, 1995.
[12] S. Oh and S. Sastry, “Tracking on a graph,” in Proceedings of [27] S. Oh, L. Schenato, P. Chen, and S. Sastry, “Tracking and
the 4th International Symposium on Information Processing in coordination of multiple agents using sensor networks: system
Sensor Networks (IPSN ’05), pp. 195–202, Los Angeles, Calif, design, algorithms and experiments,” Proceedings of the IEEE,
USA, April 2005. vol. 95, no. 1, pp. 234–254, 2007.
[13] R. Brooks, D. Friedlander, J. Koch, and S. Phoha, “Tracking [28] T. Kurien, “Issues in the design of practical multitarget track-
multiple targets with self-organizing distributed ground sen- ing algorithms,” in Multitarget-Multisensor Tracking: Advanced
sors,” Journal of Parallel and Distributed Computing, vol. 64, Applications, Y. Bar-Shalom, Ed., Artech House, Norwood,
no. 7, pp. 875–884, 2004. Mass, USA, 1990.
[14] C. Sharp, S. Schaffert, A. Woo et al., “Design and implemen- [29] J. Hill, M. Horton, R. Kling, and L. Krishnamurthy, “The plat-
tation of a sensor network system for vehicle tracking and forms enabling wireless sensor networks,” Communications of
autonomous interception,” in Proceedings of the 2nd European the ACM, vol. 47, no. 6, pp. 41–46, 2004.
Workshop onWireless Sensor Networks (EWSN ’05), pp. 93–107, [30] S. Oh, I. Hwang, and S. Sastry, “Distributed multitarget track-
February 2005. ing and identity management,” Journal of Guidance, Control,
[15] J. Aslam, Z. Butler, F. Constantin, V. Crespi, G. Cybenko, and and Dynamics, vol. 31, no. 1, pp. 12–29, 2008.
D. Rus, “Tracking a moving object with a binary sensor net- [31] S. Meguerdichian, F. Koushanfar, G. Qu, and M. Potkonjak,
work,” in Proceedings of the ACM 1st International Conference “Exposure in wireless ad-hoc sensor networks,” in Proceedings
on Embedded Networked Sensor Systems (SenSys ’03), pp. 150– of the 7th Annual International Conference on Mobile Comput-
161, November 2003. ing and Networking, pp. 139–150, July 2001.
[16] W. Chen, J. Hou, and L. Sha, “Dynamic clustering for acoustic
[32] H. Pasula, S. Russell, M. Ostland, and Y. Ritov, “Tracking many
target tracking in wireless sensor networks,” in Proceedings of
objects with many sensors,” in Proceedings of the International
the 11th IEEE International Conference on Network Protocols,
Joint Conference on Artificial Intelligence, Stockholm, Sweden,
pp. 284–294, November 2003.
July 1999.
[17] M. Coates, “Distributed particle filters for sensor networks,” in
[33] G. Roberts, “Markov chain concepts related to sampling algo-
Proceedings of the 3rd International Symposium on Information
rithms,” in Markov Chain Monte Carlo in Practice, W. Gilks,
Processing in Sensor Networks (IPSN ’04), pp. 99–107, April
S. Richardson, and D. Spiegelhalter, Eds., Interdisciplinary
2004.
Statistics, Chapman and Hall, New York, NY, USA, 1996.
[18] J. Liu, J. Liu, J. Reich, P. Cheung, and F. Zhao, “Distributed
[34] K. Whitehouse, F. Jiang, A. Woo, C. Karlof, and D. Culler,
group management for track initiation and maintenance in
“Sensor field localization: a deployment and empirical anal-
target localization applications,” in Proceedings of the 2nd
ysis,” Tech. Rep. UCB//CSD-04-1349, University of California,
workshop on Information Processing in Sensor Networks, pp.
Berkeley, Calif, usa, 2004.
113–128, April 2003.
[19] F. Zhao, J. Liu, J. Liu, L. Guibas, and J. Reich, “Collaborative [35] Berkeley, “Mica2dot hardware design files,” 2003, motes man-
signal and information processing: an information-directed ufactured by Crossbow Technology Inc., http://www.tinyos
approach,” Proceedings of the IEEE, vol. 91, no. 8, pp. 1199– .net/scoop/special/hardware.
1209, 2003. [36] J. Hill, R. Szewczyk, A. Woo, S. Hollar, D. Culler, and K. Pister,
[20] J. Liu, J. Reich, and F. Zhao, “Collaborative in-network pro- “System architecture directions for networked sensors,” in Pro-
cessing for target tracking,” Eurasip Journal on Applied Signal ceedings of the 9th Internatinal Conference Architectural Support
Processing, vol. 2003, no. 4, pp. 378–391, 2003. for Programming Languages and Operating Systems (ASPLOS-
[21] D. McErlean and S. Narayanan, “Distributed detection and IX ’00), pp. 93–104, Cambridge, Mass, USA, November 2000.
tracking in sensor networks,” in Proceedings of the 36th Asilo- [37] S. Bergbreiter and K. S. J. Pister, “Cotsbots: an off-the-shelf
mar Conference on Signals Systems and Computers, pp. 1174– platform for distruted robotics,” in Proceedings of the IEEE/RSJ
1178, November 2002. International Conference on Intelligent Robots and Systems
[22] J. Shin, L. Guibas, and F. Zhao, “A distributed algorithm for (IROS ’03), pp. 27–31, Las Vegas, Nev, USA, October 2003.
managing multi-target identities in wireless ad-hoc sensor [38] Honeywell, “1- and 2- axis magnetic sensors,” http://www.ssec
networks,” in Proceedings of the 2nd workshop on Information .honeywell.com/magnetic/datasheets/hmc1001-2&1021-2
Processing in Sensor Networks, pp. 223–238, April 2003. .pdf.
[23] J. Liu, M. Chu, J. Liu, J. Reich, and F. Zhao, “Distributed state [39] A. Woo, T. Tong, and D. Culler, “Taming the underlying chal-
representation for tracking problems in sensor networks,” in lenges of reliable multihop routing in sensor networks,” in
Proceedings of the 3rd International Symposium on Information Proceedings of the 1st International Conference on Embedded
Processing in Sensor Networks (IPSN ’04), pp. 234–242, April Networked Sensor Systems (SenSys ’03), pp. 14–27, Los Angeles,
2004. Calif, USA, November 2003.
[24] M. Chu, S. Mitter, and F. Zhao, “Distributed multiple target [40] M. Kushwaha, S. Oh, I. Amundson, X. Koutsoukos, and A.
tracking and data association in ad hoc sensor networks,” in Ledeczi, “Target tracking in urban environments using audio-
Proceedings of the 6th International Conference on Information video signal processing in heterogeneous wireless sensor net-
Fusion, pp. 447–454, July 2004. works,” in Proceedings of the 42nd Asilomar Conference on
[25] J. B. Collins and J. K. Uhlmann, “Efficient gating in data asso- Signals, Systems and Computers (ASILOMAR ’08), pp. 1606–
ciation with multivariate Gaussian distributed states,” IEEE 1610, Pacific Grove, Calif, USA, October 2008.
Transactions on Aerospace and Electronic Systems, vol. 28, no. [41] M. Meingast, M. Kushwaha, S. Oh, X. Koutsoukos, A. Ledeczi,
3, pp. 909–916, 1992. and S. Sastry, “Fusion-based localization for a heterogeneous
16 International Journal of Distributed Sensor Networks