A Survey of Localization in Wireless Sensor Network

Abstract

Localization is one of the key techniques in wireless sensor network. The location estimation methods can be classified into target/source localization and node self-localization. In target localization, we mainly introduce the energy-based method. Then we investigate the node self-localization methods. Since the widespread adoption of the wireless sensor network, the localization methods are different in various applications. And there are several challenges in some special scenarios. In this paper, we present a comprehensive survey of these challenges: localization in non-line-of-sight, node selection criteria for localization in energy-constrained network, scheduling the sensor node to optimize the tradeoff between localization performance and energy consumption, cooperative node localization, and localization algorithm in heterogeneous network. Finally, we introduce the evaluation criteria for localization in wireless sensor network.

Full text

Hindawi Publishing Corporation
International Journal of Distributed Sensor Networks
Volume 2012, Article ID 962523, 12 pages
doi:10.1155/2012/962523
Review A rticle
A Survey of Localization in Wireless Sensor Network
Long Cheng,
1
Chengdong Wu,
1
Yunzhou Zhang,
1
Hao Wu,
2
Mengxin Li,
3
and Carsten Maple
4
1
College of Information Science and Engineering, Northeastern University, Shenyang 110819, China
2
Faculty of Engineeri ng, University of Sydney, Sydney, 2006 NSW, Australia
3
Faculty of Information and Control Engineering, Shenyang Jianzhu University, Shenyang 110168, China
4
Department of Computer Science and Technology, University of Bedfordshire, Luton LU1 3JU, UK
Correspondence should be addressed to Long Cheng, chenglong2000
0@yahoo.com.cn
Received 18 September 2012; Accepted 16 November 2012
Academic Editor: Wei Meng
Copyright © 2012 Long Cheng et al. 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.
Localization is one of the key techniques in wireless sensor network. The location estimation methods can be classified into
target/source localization and node self-localization. In target localization, we mainly int roduce the energy-based method. Then
we investigate the node self-localization methods. Since the widespread adoption of the wireless sensor network, the localization
methods are different in various applications. And there are several challenges in some special scenarios. In this paper, we
present a comprehensive survey of these challenges: localization in non-line-of-sight, node selection criteria for localization in
energy-constrained network, scheduling the sensor node to optimize the tradeoff between localization performance and energy
consumption, cooperative node localization, and localization algorithm in heterogeneous network. Finally, we introduce the
evaluation criteria for localization in wireless sensor network.
1. Introduction
Due to the availability of such low energy cost sensors,
microprocessor, and radio frequency circuitry for informa-
tion transmission, there is a wide and rapid diffusion of
wireless sensor network (WSN). Wireless sensor networks
that consist of thousands of low-cost sensor nodes have
been used in many promising applications such as health
surveillance, battle field surveillance, and environmental
monitoring. Localization is one of the most important
subjects because the location information is typically useful
for coverage, deployment, routing, location service, target
tracking, and rescue [1]. Hence, location estimation is
a significant technical challenge for the researchers. And
localization is one of the key techniques in WSN.
The sensor nodes are randomly deployed by the vehicle
robots or aircrafts. While the Global Positioning System
(GPS) is one of the most popular positioning technologies
which is widely accessible, the weakness of high cost and
energy consuming makes it different to install in every node.
In order to reduce the energy consumption and cost, only
a few of nodes which are called beacon nodes contain the
GPS modules. The rest of nodes could obtain their locations
through localization method. The process of estimating the
unknown node position within the network is referred to
as node self-localization. And WSN is composed of a large
number of inexpensive nodes that are densely deployed in
a region of interests to measure certain phenomenon. The
primary objective is to determine the location of the target.
As shown in Figure 1, we classify the localization method
into target/source localization and node self-localization.
And the target localization can be further classified into four
categories: single-target localization in WSN, multiple-target
localization in WSN, single-target localization in wireless
binary sensor network (WBSN), and multiple-target local-
ization in WBSN. And node self-localization can be classified
into two categories: range-based localization and range-
free localization. The former method uses the measured
the distance/angle to estimate the location. And the latter
method uses the connectivity or pattern matching method to
estimate the location. We will present the localization
method in some special scenarios and finally introduce the
evaluation criteria for localization in WSN.

2 International Journal of Distributed Sensor Networks
Localization in
wireless sensor
network
Target/source
localization
Node self-
localization
Single-target
localization
in WSN
Multiple-
target
localization
in WSN
Single-target
localization
in WBSN
Multiple-
target
localization
in WBSN
Range-based
localization
Range-free
localization
Pattern
matching
localization
Hop-count-
based
localization
TOA TDOA RSSI AOA
Figure 1: Localization methods taxonomy.
2. Target/Source Localization
2.1. Single-Target/Source Localization in Wireless Sensor Net-
work. The source localization methods have a wide range
of possible applications. The outdoor application includes
vehicle or aircraft localization. In an indoor environment,
this method could track the human speakers. In underwater
environment, it can be used to locate the large sea animals
and ships. There are se veral ways to estimate the source loca-
tion: energy-based, angle of arrival (AOA) [2], time differ-
ence of arrival ( TDOA) [3–6]. As an inexpensive approach,
energy-based method is an attractive method because
it requires low hardware configuration. In this survey,
we focus on the energy-based source localization.
Single-source localization can be further divided into: en-
ergy decay model-based localization algorithm and model-
independent localization algorithms.
(1) Decay Model-Based Localization Algorithm. Equation (1)
shows the decay model in [7–9]. The received signal strength
at ith sensor during time interval t can be written as
y
i
(
t
)
= g
i
S
(
t
)
d
2
ik
(
t
)
+ n
i
(
t
)
,
(1)
where g
i
represents the gain factor of the ith sensor. We
assume that g
i
= 1. S(t) is the signal energy at 1 meter away.
And d
ik
is the Euclidean distance between the ith sensor and
the source. In addition n
i
is the measurement noise modeled
as zero mean w hite Gaussian with var iance σ
2
i
,namely,n
i
∼
N (0, σ
2
i
).
Although this energy decay model appears quite sim-
plistic, it is the one commonly used in the literature. Since
the objective function of single-source localization method
has multiple local optima and saddle points [7], the authors
formulated the problem as a convex feasibility problem
and proposed a distributed version of the projection onto
convex sets method. A weighted nonlinear least squares and
weighted linear least squares methods [8] were proposed
to estimate the location of the target. In [9], the authors
proposed normalized incremental subgradient algorithm to
solve the energy-based sensor network source localization
problem where the decay factor of the energy decay method
is unknown.
Unlike the signal models in [7–9], the authors derived
a more generalized statistical model [10] for energy obser-
vation. And a weighted direct/one-step least-squares-based
algorithm was investigated to reduce the computational
complexity. And in comparison with quadratic elimination
method, these methods were amenable to a correction
technique which incorporates the dependence of unknown
parameters leading to further performance gains. This
method offered a good balance between the localization
performance and computational complexity. Energy ratio
formulation [11] was an alternative approach that is inde-
pendent of the source energy S(t). This was accomplished
by taking ratios of the energy reading of a pair of sensors in
the noise-free case. In [12], the authors proposed an energy
aware source localization method to reduce the energy
consumption in localization.
(2) Model-Independent Methods. A kernel averaging ap-
proach [13] which needs not information about energy decay
model was proposed. In [14], a novel model-independent
localization method was proposed. Since the nodes with
higher received signal strength measurement were closer to
the source, a distributed sorting algorithm is employed. If
the sensor nodes know their rank, the required distance
estimates are obtained as the expected value of the respective
probability density functions. Finally, the projection onto
convex sets (POCS) method was used to estimate the location
of the source.
2.2. Multiple-Target Localization in Wireless Sensor Network.
Many works investigate the single-target localization. How-
ever, very limited papers investigate the multiple-target
localization. Most of the works are based on the maximum
likelihood estimator. The details of the maximum likelihood
estimator are as follows.

International Journal of Distributed Sensor Networks 3
The received signal strength at ith sensor during time in-
terval t can be written as
y
i
(
t
)
= g
i
K

k=1
S
k
(
t
)
d
α
ik
(
t
)
+ ε
i
(
t
)
,
(2)
where d
ik
(t) is the distance between the ith sensor and the
kth source. K is the number of the sources. g
i
is the gain of
ith sensor. ε
i
(t) is random variable with mean μ
i
and variance
σ
2
i
. S
k
(t) is the signal energy at 1 meter away for kth source.
α is the attenuation exponent.
We define the following matrix notations as follows:
Y
=

y
1
− μ
1
σ
1
, ...,
y
N
− μ
N
σ
N

T
,
G
= diag

1
σ
1
, ...,
1
σ
N

,
S
=
[
S
1
, S
2
, ..., S
K
]
T
,
D
=
⎡
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎣
g
1
d
2
11
g
1
d
2
12
, ...,
g
1
d
2
1K
g
2
d
2
21
g
2
d
2
22
, ...,
g
2
d
2
2K
.
.
.
.
.
. ...
.
.
.
g
N
d
2
N1
g
N
d
2
N2
, ...,
g
N
d
2
NK
⎤
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎦
, ε =
[
ε
1
, ε
2
, ..., ε
N
]
N
.
(3)
Using these notations, (2)canberepresentedas
Y
= GDS + ε = HS + ε,
(4)
where H
= GD.
So the joint probability density function (4)canbe
expressed as
f
(
Y
| θ
)
=
(
2π
)
−N/2
exp

−
1
2
(
Y
− HS
)
T
(
Y
− HS
)

,(5)
where θ
= [r
1
, r
2
, ..., r
K
; S
1
, S
2
, ..., S
K
]
T
.
The maximum likelihood estimation is equivalent to mini-
mizing the follow ing function:
L
(
θ
)
=
(
Y
− HS
)
T
(
Y
− HS
)
=Y − HS
2
.
(6)
We can obtain the maximum likelihood parameter estima-
tion of θ by minimizing L(θ).
To minimize L(θ), we should take the following opera-
tion:
∂L
(
θ
)
∂S
i
= 0.
(7)
This condition leads to the following relation:
S
= H
∇
H,whereH
∇
is the pseudoinverse of the matrix
H.
So we get the modified cost function:
arg min L

(
θ
)
=



Y − HH
∇
Y



2
.
(8)
And a multiresolution (MR) search and the expectation
maximization (EM) method [15] were proposed to solve (8).
An efficient EM algorithm [16] was proposed to improve
the estimation accuracy and avoid trapping into local
optima through the effective sequential dominant-source
initialization and incremental search schemes. An alternating
projection [17] algorithm was proposed to decompose the
multiple-source localization into a number of simpler, yet
also nonconvex, optimization steps. This method could
decrease the computation complexity.
2.3. Single-Target/Source Localization in Wireless Binary Se n-
sor Network. Most of the source localization methods are
focused on the measured signal strength; that is, the fusion
center knows the measurements of the nodes. In order
to obtain the measurements, the node needs the complex
calculating process. The above methods require transmission
of a large amount of data from sensors which may not
be feasible under communication constraints. The binary
sensors sense signals (infrared, acoustic, light, etc.) from
their vicinity, and they only become active by transmitting
a signal if the strength of the sensed sig nal is above a certain
threshold. The binary sensor only makes a binary decision
(detection or nondetection) regarding the measurement, and
consequently, only its ID needs to be sent to the fusion center
when it detects the target, otherwise it remains silent. So
the binary sensor is a low-power and bandwidth-efficient
solution for wireless sensor network.
Limited papers investigate the source localization in
binary sensor network. And previous works have been
proposed to try to estimate the location of the single source in
wireless binary sensor network (WBSN). In [18], the authors
proposed a maximum likelihood source location estimator
in WBSN. A low complexity source localization method [19]
which is based on the intersection of detection areas of
sensors was introduced in noisy binary sensor networks. A
subtract on negative add on positive (SNAP) [20] algorithm
was proposed to identify the source location using the binary
sensor networks. This is a fault-tolerant algorithm that is
slightly less accurate but it is computationally less demanding
in comparison with maximum likelihood estimation. In
[21], the authors proposed a trust index based subtract on
negative add on positive (TISNAP) method to improve the
accuracy of localization for multiple event source localiza-
tion. This a lgorithm reduces the impac t of faulty nodes on
the source localization by decreasing their trust index. And
the TISNAP algorithm assumed that the distance between
any two sources is far enough; that is, the node is influenced
by only one source initially. So the localization process is
similar to the single-source localization process. However,
all of the previous works mainly focus on single-source
localization. Fewer papers investigate the multiple-source
localization in WBSN.
3. Node Self-Localization
3.1. Range-Based Localization. The classic methods to esti-
mate the indoor location are time of arrival (TOA), time

4 International Journal of Distributed Sensor Networks
difference of arrival (TDOA), angle of arrival (AOA), and
received signal strength (RSS). TOA method measures t ravel
times of signals between nodes. TDOA method locates
by measuring the signals’ arr ival time difference between
anchor nodes and unknown node. It is able to achieve
high ranging accuracy, but requires extra hardware and
consumes more energy. As an inexpensive approach, RSS has
established the mathematical model on the basis of path loss
attenuation with distance [22, 23], and it requires relatively
low configuration and energy. We can obtain the distance
between the beacon node and unknown node through the
above three measurement methods. We set the position of
beacon node is
(x
1
, y
1
), ...,(x
N
, y
N
),  and the position of
unknown node is X
= [x, y]
T
.

d
i
is the estimated distance
between ith beacon node and unknown node. We can obtain
the coordinate matrix of the unknown node as follows:
X
=

A
T
A

−1
A
T
B,
A
= 2
⎡
⎢
⎢
⎢
⎢
⎣
(
x
1
− x
2
)

y
1
− y
2

(
x
1
− x
3
)

y
1
− y
3

.
.
.
.
.
.
(
x
1
− x
N−1
)

y
1
− y
N−1

⎤
⎥
⎥
⎥
⎥
⎦
,
B
=
⎡
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎣

d
2
2
−

d
2
1
−

x
2
2
+ y
2
2

+

x
2
1
+ y
2
1


d
2
3
−

d
2
1
−

x
2
3
+ y
2
3

+

x
2
1
+ y
2
1

.
.
.

d
2
N
−1
−

d
2
1
−

x
2
N
−1
+ y
2
N
−1

+

x
2
1
+ y
2
1

⎤
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎦
.
(9)
The angles between unknown node and a number of
anchor nodes are used in the AOA method to estimate the
location. This method needs the antenna array which is an
expensive solution for low-cost sensor node.
3.2. Range-Free Localization
3.2.1. Hop-Count-Based Localization. As range-free posi-
tioning system, DV-Hop is the typical representation. It
does not need to measure the absolute distance between the
beacon node and unknow n node. It uses the average hop
distance to approximate the actual distances and reduces
the hardware requirements. It is easy to implement and
applicable to large network. But the positioning error is also
correspondingly increased.
The positioning process of DV-Hop is divided into
three stages: information broadcast, distance calculation, and
position estimation. In information broadcast stage, the
beacon nodes broadcast their location information package
which includes hop count and is initialized to zero for their
neighbors. The receiver records the minimal hop of each
beacon nodes and ignores the larger hop for the same beacon
nodes. Then the receiver increases the hop count by 1 and
transmits it to neighbor nodes. All the nodes in a network
can record the minimal hop counts of each beacon nodes.
In distance calculation stage, according to the position of
the beacon node and hop count, each beacon node uses the
following equation to estimate the actual distance of every
hop:
HopSize
i
=

j
/
= i


x
i
− x
j

2
+

y
i
− y
j

2

j
/
= i
h
j
,
(10)
where (x
i
, y
i
)and(x
j
, y
j
) are the coordinates of beacon
nodes i and j,respectively.h
j
is the hop count between
the beacon nodes. Then, beacon nodes will calculate the
average distance and broadcast the information to network.
The unknown nodes only record the first average distance
and then transmit it to neighbor nodes. Finally, the unknown
node calculates its location through (9). In order to improve
the localization accuracy, the improved algorithm mainly
focuses on the following several aspects: average hop distance
between beacon nodes, deployment of the beacon nodes, and
node information.
(1) Average Hop Distance between Beacon Nodes. In the
randomly deployed node density and connectiv ity network,
Wang et al. [24] proposed a hop progress analytical model
to estimate the optimal path distance between any pair of
sensor nodes in the network. And it derived an expected
hop progress and hop counts estimation method. A range-
free localization algorithm (LAEP) which is using the trilat-
eration techniques and the expec ted hop progress analytical
results is proposed. Unlike the DV-Hop method, the LAEP
broadcasts the anchor coordinated and the corresponding
estimated distance to each sensor at the same time; therefore,
it can dramatically reduce the network traffic and the com-
munication delay. Wang only considers the node’s receipt
of beacon on a line to the utmost extent. Xu et al. [25]
proposed a mobile anchor node localization method that is
based on Archimedes curve. It takes communication path
as curve spread. It avoids the error caused by large straight
line dissemination and improves the precision. Lee et al. [26]
proposed a robust weighted algorithm which is based on
DV-Hop algorithm to calculate the average hop distances
between unknown nodes and anchor nodes. It applies to
most topological st ructure networks and reduces the location
error. In the same way, Zhang et al. [27] improved the
average hop distance based on minimum mean square error
standard, which reduced positioning error. Lee et al. [28]
used Karush-Kuhn-Tucker (KKT) standards and Lagrange’s
mean value theorem to correct the average hop distance error
and improve location accuracy.
(2) Dep loyment of the Beacon Nodes. According to the ideal
or regular node deployment scheme, the modified DV-Hop
method is improved. Zheng et al. [29] firstly derived a beacon
nodes deployment strategy that deploys a beacon node in the
center of the area and other nodes are e qually placed in the
circle whose center is the center of the area and radius is half
of the length of the area. Based on this deployment strategy,
an accurate long-range DV-Hop algorithm is proposed. This
method is adapted to large-scale network. Lee et al. [30]
put forward a quadratic programming method to optimize

International Journal of Distributed Sensor Networks 5
adjacent distance mapping. And it can be applied to the
isotropic and anisotropic network.
(3) Node Information. Some modified methods were pro-
posed through the neighbor node information such as
anchor information and relationship between node and
anchor or topology structure to improve the DV-Hop
method. Zhong and He [31] proposed a proximity metric
called RSD (regulated signature distance) to capture the
distance relationships among 1-hop neighboring nodes.
This method can be conveniently applied as a transpar-
ent supporting layer for state-of-the-art connectivity-based
localization solutions to achieve better accuracy. He et
al. [32] proposed a spring swarm localization algorithm
(SSLA) which uses the network topology information and
a small amount of anchor node location information to
calculate unknown nodes position. Chen et al. [33] used
the information of neighbor node to make anchor node
communication range to be gradient to improve accuracy.
Lim and Hou [34] addressed the issue of localization in
anisotropic sensor networks. And a linear mapping method
is proposed to characterize anisotropic features. It projects
one embedding space built upon proximity measures into
geographic distance space by using the truncated singular
value decomposition (SVD) pseudoinverse technique. This
method is different from MDS-Map method and owns
higher accuracy than MDS-Map algorithm and the other
expanded MDS-Map algorithm.
(4) Comprehensive Improvement Method. In addition to the
above aspects, Brida et al. [35] used DV-Hop algorithm,
DV-distance, DV-Euclidean algorithm, the constraint, and
iteration condition that are to be added to select reference
node. Then it used trilateration to find possible unknown
node area, the estimated center of area as the final position.
This algorithm could reduce the network energy consump-
tion and improves localization accuracy. Chia-Ho [36]put
forward a distributed, range-free localization algorithm. In
this method, the mobile beacon node with directive antenna
is used to supply the location information for the unknown
node. Tan et al. [37] exploited acoustic communication to
further research underwater range-free algorithm, especially
proposing future prospect and development trend in view of
the characteristic of underwater acoustic channel.
3.2.2. Pattern Matching Method. Pattern matching localiza-
tion, also called map-based or Fingerprint algorithm, is
one of the most viable solutions for range-free localization
methods re cently. The fingerprint localization involves two
phases. During the first phase, the received signals at selected
locations are recorded in an offline database called radio
map. Then, the second phase, it works at the online state. The
pattern matching algorithms are used to infer the location
of unknown node by matching the current observed signal
features to the prerecorded values on the map [38, 39].
Fang et al. [40] proposed a novel method to extrac t
the feature of robust signals to efficiently mitigate the
multipath effect. This method enhances the robustness under
a multipath fading condition and is commonly used for
the indoor environments. Swangmuang and Krishnamur thy
[41] presented a new analytical model that applies prox-
imity graphs for approximating the probability distribution
of error distance, which recorded a location fingerprint
database by using the received signals. In addition, there are
many problems under the indoor positioning environment
as following: how to capture the character of propagation
signal in the complex dynamic environment and how to
accommodate the receiver gain difference of different mobile
devi ces and so on. Wang et al. [42] solved these problems by
modeling them as common mode noise and then developed
a location algorithm based on a novel differential radio
map. Gogolak et al. [43] proposed fingerprint localization
methodology based on neural network which is applied in
the real experimental indoor environment. It provided the
necessary measurement results to the fingerprint localiza-
tion.
4. Localization in Some Special Scenarios
Current and potential applications of sensor networks may
be quite different. The scale of the network in these
applications may be small or large, and the environments
may be different. So the traditional localization methods are
not suitable for the special scenarios. And there are some
challenges for locating sensor nodes that need to be solved. In
this survey, we mainly describe the following four challenges.
The first challenge is NLOS (non-line-of-sight) r anging error
problem. The direct path from the unknown node to the
beacon is blocked by obstacles in wireless sensor network;
the signal measurements include an error due to the excess
path traveled because of the reflection of acoustic signal,
which is termed as the NLOS error. The NLOS error results
in the large location estimation error. The second challenge is
the energy consumption and localization accuracy problem.
Since the sensor node is powered by battery, the node may fail
due to the depletion of energy. So the energy consumption
is critical for the localization problem. It contains the node
selection, tradeoff between localization performance and
energy consumption, and node resource management. Since
the unreliable hardware and complicate communication
environment, the received information may be unreliable.
Therefore, the third challenge is corporative localization.
And the fourth challenge is the localization in heterogeneous
sensor network.
4.1. Localization in NLOS Scenario
4.1.1. NLOS Identification/Classification. As shown in Fig-
ure 2, there may be no direct path from the beacon node
to the unknown node in complicated environment. Because
of the reflection and diffraction, the signal which is used for
distance measurement can reflect and bound off multiple
surfaces before arriving at the receiver. So the signal may
actually travel excess path lengths and the direct path is
blocked. The signal measurements include an error due to
the excess path traveled w hich is termed as the NLOS error.

6 International Journal of Distributed Sensor Networks
Obstacle
Obstacle
Obstacle
Beacon node
Unknown node
NLOS
propagation
LOS
propagation
Figure 2: Example of localization in LOS/NLOS environments.
TheNLOSproblemhasbeenstudiedin[44] and it
was reported that the NLOS error was quite common in all
environments except for rural areas. The large location esti-
mation error will occur in NLOS environment. Accordingly,
NLOS identification is particularly significant in localization.
The problem of NLOS identification is essentially a detection
problem. There are two main approaches to solve the NLOS
errors: parametric methods and nonparametric methods. In
this section, we give a brief overview of some key researches
in this area.
Wylie and Holtzman [45] proposed a method based
on parameter hypothesis test which determines the mea-
surements w hether belong to NLOS by comparing the
NLOS variance with the LOS variance. It has a simple
criterion but needs the detailed environment parameter and
a prior knowledge. Borras et al. [46]investigatedNLOS
identification by using binary hypothesis test and generalized
likelihood ratio for identifying NLOS error. It proposed a
proper decision criteria and its premise is that the NLOS
error is Gaussian distributed with a large variance. Based on
Wylie-Holtzman algorithm, Mazuelas et al. [47] proposed an
improvement by using NLOS ratio estimation and it will be
able to correct the NLOS measurements from the previous
knowledge of this ratio.
The aforementioned methods need a mount of priori
knowledge and historical data. Chan et al. [48]proposeda
residual test (RT) method to overcome the shortage which
needs lots of priori knowledge. It determines the measure-
ments whether belong to NLOS by measuring the samples
appropriate to central Chi-distribution. The principle of this
method is that if all measurements are LOS, and if the
localization technique gives maximum likelihood estimates,
then the residuals, normalized by the Cramer Rao Lower
Bound (CRLB), will have a central χ
2
distribution. And if
the measurements contain the NLOS error, the distribution
is noncentral χ
2
distribution. Venkatraman and Caffery Jr.
[49] investigated NLOS identification for moving targets
by using a time series of range measurements. Gezici et
al. [50] proposed a nonparameter-based hypothesis test
method which used a distance metric between a known
measurement error distribution and a nonparametrically
estimated distance measurement distribution. Yu and Guo
[51] also proposed a nonparameter-based hypothesis test
method by using generalized likelihood ratio to establish
the relationship between LOS and NLOS. Then it used
Neyman-Pearson (NP) test method for NLOS identification.
A sequential probability ratio test [52] which is tolerant to
the parameters fluctuations is employed to identify whether
the measurement contains the non-line-of-sight (NLOS)
errors.
4.1.2. NLOS Mitigation. Because of the existence of the non-
ideal channel condition and non-line-of-sight transmission
between the unknown nodes and beacon nodes, NLOS error
mitigation has become a key technology and hotspot in
the research about location estimation in wireless sensor
network.
The first way attempts to identify the propagation
conditions (LOS or NLOS) and then eliminate the mea-
surements in NLOS; they only use the measurements in
LOS to locate the unknown node. The propagation model-
based method [53, 54] either directly employs the existing
propagation models or empirically develops a model based
on experimental results. The second way uses al l NLOS and
LOS measurements to estimate the location, but provides
weighting or sealing to minimize the effects of the NLOS
contributions. The weighting is determined by either the
position geometry and beacon nodes layout or the residuals
(fitting errors) of individual beacon node. The Taylor
series linearization [55] (TS-LS), a widely used localization
algorithm, should have the prior information of the error
statistics which can be used to determine the weights. Chen
[56] develops an algorithm to mitigate the NLOS er rors by
residual weighting when the range measurements corrupted
by NLOS errors are not identifiable. The hypothesis testing
[57] is employed to detect whether the environment is NLOS
or LOS along with time of arrival ( TOA) and received signal
strength (RSS) measurements. And then an extended Kalman
filter is used to nonlinear estimation.
In a scattering environment, most of the propagation
paths between the unknown nodes and beacon nodes are
NLOS; the constrained optimization techniques are used to
reduce NLOS errors [58].Theneuralnetworkisemployed
to predict the NLOS error [59]; Kalman filters [60]and
modified two-stage Kalman filter [61] are used to correct
NLOS measurements.
All the positioning algorithms in NLOS environment
focused more on the cellular network. These methods could
be used in some scenarios for wireless sensor network
localization. Fewer methods investigate the NLOS mitigation
algorithm for Wireless sensor network.
4.2. Node Selection Criteria for Localization in Energy-
Constrained Network. Due to the limited power of sen-
sor node and hostile deployment environment, the node
selection in WSN is different from the traditional node
selection in traditional w ireless network. If all the sensor
nodes are used at the same time to executive localization

International Journal of Distributed Sensor Networks 7
task without selection, although the energy consumption of
nodes selection are saved, but at this time the repeatability of
the received information would be quite larger. If the nodes
are random selected to executive the localization task, the
algorithm is simple and the extra overhead can be ignored,
but localization accuracy is low in this case. Obviously,
this method cannot satisfy the user’s requirements for the
accurate localization, the unbalance of energy consumption
will appear, and some nodes may fail due to the depletion
of energy. This may affect the network connectivity and
may result in losing the sensed data. These characteristics
of WSN determine selection method which is different from
traditional network. Therefore, it is necessary to investigate
nodes’ selection mechanism in WSN.
The primary algorithm makes decision with global infor-
mation [62]; this method minimized the expected filtered
mean-squared position error for a given number of active
nodes by using a global knowledge of all node locations. This
algorithm needs the positions of nodes and broadcasts them
to all nodes and a lot of data communication; therefore, it
only can be applied to small networks. Based on the former
algorithm, a local selection strategy is investigated [63].
This method determines whether or not that node should
be active by only incorporating geometrical knowledge
of itself and the active set of nodes from the previous
information. Based on this approach, the researchers have
also investigated other strategies, such as the least square
method, Bayes probability method [64, 65]. Furthermore,
in order to narrow the scope and scale of selected nodes,
researchers proposed a method which combines the track
and the current state of the robot. Zhang and Cao [66]
proposed a multinode cooperation dynamic tree algorithm.
This method ensured that spanning tree has low energy
consumption and high information content by increasing
and decreasing the number of the nodes dynamically. But
this method still had some disadvantages: the root node
needs data fusion and the new node needs to b e calculated,
and the consumption of energy is quite higher. Yang et al.
[67] proposed online prediction based on particle filter and
estimate the probability distribution of the target state under
the Bayes framework. This method realized the optimal
selection of the node sequence and introduced a shortest
path algorithm to reduce the information transmission.
Hamouda and Phillips [68] proposed a method which
employs the moving speed of the mobile robot to improve
the localization accuracy and consistency.
The signal shielding and multipath interference make the
channel parameters become too complex to definite error
factor. Bel et al. [69] proposed two selection principles to
reduce the number of active nodes, and the nodes with
accurate measured value (RSSI value larger than specified
threshold) are selected. This method is effective to balance
the accuracy and energy consumption and is suitable for
the WSN which is hardware resource constrained. Zhao
and Nehorai [70] used the Cramer-Rao equation to select
the next node to participate in positioning. Because of the
complexity of the observed model and the non-Gaussian
noise, it is hard to get the optimal solution of the problem.
4.3. Scheduling the Se nsor Node to Optimize the Tradeoff
between Localization Performance and Energy Consumption.
A typical sensor network consists of a large number of small
sensors which are deployed randomly. However, a sensor
node has limited resources because of battery power and
small memory. Therefore, nodes’ resource management is
compulsory. In typical sensor network applications, nodes
are deployed in an unattended environment such as disaster
management, habitat monitoring, industrial process control,
and object tracking. Enormous event data will be generated
for a long sensing time in WSN. Hence, by the methods
of nodes resource management, effective usage of the vast
amount of data is crucial. In the meanwhile, the scalability
of both energy and spatial dimensions in distributed sensor
network is a key issue. Sensor networks must track various
phenomena at the same time and work within limited
communication bandwidth, energy, and processing speed.
Therefore, it is critical to distribute the workload across
only the “relevant” sensors equally and leave other sensors
available for other things. These characteristics of WSN
determine the importance of nodes resource management.
Energy consumption is one of the most important issues
in recent years. Ren and Meng [71] proposed a localization
algorithm based on particle filtering for sensor networks.
Assisted by multiple-transmit-power information, it outper-
forms the existing algorithms that do not utilize multiple-
power information. You et al. [72] proposed a specified
positional error tolerance, the sensor-enhanced and energy-
efficient adaptive localization system in an application. This
localization system dynamically sets sleep time for the nodes
and adapting the sampling rate of target’s mobility level.
However, the process of error estimation dynamically relies
on several factors in the specific environment. Gribben
et al. [73] proposed a scheduling algorithm that selects a
subset of active beacon nodes to be used in localization. It
served to reduce the message overhead, increased network
lifetime, and improved localization accuracy in dense mobile
networks. However, maximizing the nodes’ sleep time is
much more energy efficient if the nodes never wake up until
the reception of wake-up messages. The a bove algorithms
have the same feature that the duty cycle of the sensor
nodesisfixedinadvance.In[74], the authors proposed
an innovative probabilistic wake-up protocol for energy-
efficient event detection in WSNs. The main idea of it is to
reduce the duty cycle of every sensor via probabilistic wake-
up through the dense deployment of sensor networks.
The problem of unique network localization and a math-
ematical topic known as rigidity theory have a strong
connection. Goldenberg et al. [75] proposed a localization
method for sparse networks by sweeping techniques. This
method is saving all possible p ositions in each position
step and pruning incompatible ones. One drawback of
sweeping method is that the possible positions could increase
exponentially as long as the number of nodes increased.
Other ty pes of localization methods are also available, such
as using multidimensional scaling [76, 77] or mobile anchors
[78, 79]. However, all the previous works tried to localize

8 International Journal of Distributed Sensor Networks
more sensor nodes in a network without guaranteeing all of
them. Khan et al. [80] introduced a localization method to
localize all nodes by the minimal number of anchor nodes.
However, they assume that the sensing range of each sensor
can be enlarged to guarantee certain triangulation, so that
three anchor nodes are enough to localize all sensors.
4.4. Cooperative Node Localization. There may be not
enough information in the concentrated network or the node
may contain the harmful information in sparse network.
There are two branches in this area: (1) access the accuracy
and reliability of the neighborhood nodes. (2) Improve
precision with the cooperation of the active and passive
nodes.
Some nodes may bring unreliable or even harmful infor-
mation [81], so it is essential to review the received infor-
mation. Tam et al. [82] employed the nearest link as reference
to review the information. When there are massive link
in dense network and positioning mainly depends on the
geometry of the neighbor node topology information, the
nearest neighbors may not correspond to the best link.
Aiming at this issue, Denis et al. [83] proposed an adaptive
method to eliminate the inefficient links, but this method has
to work with neighbor node information, and the method
cannot effectively reduce the number of packet effectively.
Therefore, Das and Wymeersch [84] put forward a kind
of distributed criterion; this method employed Cramer-Rao
limit as identifiable parameters to identify the links. This
method could avoid the invalid neighbor node links and
unreliable transmission; thus, it can effectively reduce the
computation time and the number of packets.
The accuracy of master-slave node cooperative local-
ization is mainly depended on the measurements accuracy
and the number of primary reference nodes (PRN, Primary
Reference Node). But in the actual application, it is difficult
to increase the number of primary reference nodes because
of the factors of energy and the complexity. Wymeersch et
al. [81]andFujiwaraetal.[85] put forward a new method:
the nodes which received the information of the target
and the primary reference nodes are termed as secondary
reference nodes (SRN, secondary Reference Node), the
SRNs participated in the localization in a passive way.
This cooperative method reduced the required number of
PRN with relatively higher localization accuracy. Gholami et
al. [86] used the maximum likelihood estimation method
to obtain the target position. The authors formulated the
localization problem into finding the intersection of the
vertex set by using geometry description. This method avoids
getting into the local optimum.
4.5. Localization Algorithm in Heterogeneous Sensor Network.
Most of the localization methods for the wireless sensor
networks are only to consider the homogeneous network.
The different kinds of the nodes such as the different
maximum communication radius and the different nodes
own the different localization mechanisms are not consid-
ered in homogeneous, so the localization methods for the
homogeneous network cannot be directly applied in the
heterogeneous wireless sensor networks.
Du et al. [87] propose a new boundary nodes localization
method by using a small number of anchor nodes. First the
boundary nodes are elected and their positions are deter-
mined. Then the location information of boundary nodes
is sent to other nodes through a small hop communication
range. Finally, other nodes estimate their locations by the
hopcountandhoprange.Theschemeusesfewerbeacon
nodes, but has much smaller localization error and standard
deviation. This method uses fewer beacon nodes, but with
a smaller location error and standard deviation. Dong et al.
[88] proposed a two-step localization method for two-tiered
hierarchical heterogeneous sensor networks. The network
consists of three types of nodes: a nchor nodes with known
locations, a few nodes equipped with both Ultrawide Band
(UWB) and RF radios, and a large number of normal sensor
nodes. The localization method works in two steps: firstly
the high-accurate ranging capability of UWB nodes is used
to estimate their location from a few anchor nodes, then,
sensor nodes estimate their locations by using UWB nodes
as anchor nodes.
Sometimes the distance between some nodes can be
measured directly, while others cannot be. Selecting a dif-
ferent positioning algorithms accord to the mutual distance
between nodes can be measured or not. Chiang et al.
[89] proposed a hybrid unified Kalman tracking (HUKT)
technique. The accuracy of tracking is based on both time
of arrival (TOA) and time diff
erence of arrival (TDOA)
measurements. This method is proposed to adaptively adjust
the weighting value between the TOA and TDOA measure-
ments. The scheme can both provide higher localization
accuracy for mobile network and adapt to environments with
insufficient signal sources.
According to the different communication radius of
the nodes, some super nodes can be deployed at some
areas with plenty communication demands to transmit the
information. Shen and Pesch [90] considered the nodes with
more power and longer communication range as the het-
erogeneous nodes and propose a heuristic relay positioning
algorithm for heterogeneous wireless sensor networks, to
achieve the sharing of resources in heterogeneous wireless
sensor network by using the relay nodes.
5. Evaluation Criteria for Localization in
Wireless Sensor Network
The localization errors are inevitable in the estimations. In
thissection,wedescribesomecommonmetrics:average
localization error, root mean square error, and geomet ric
mean error. And the Euclidean distance and Manhattan
distance are two widely used metr ics that are computed
considering a two-dimensional coordinate system [91]. The
Euclidean distance is defined to be the shortest distance
between two coordinates. The Manhattan distance is defined
to be the distance between two coordinates measured along

International Journal of Distributed Sensor Networks 9
the axes at the right angles. The metrics are described as
follows.
(1) Average Localization Error. The average localization er ror
for Euclidean distance can be computed as follows:
error
=
1
N
t
N
t

i=1

(
x
i
− x
)
2
+


y
i
− y

2
,
(11)
where N
t
is the number of trails. (x, y) is the true location
of the unknown node or source. (
x
i
, y
i
) is the estimated
location.
The average localization error for Manhattan distance
canbecomputedasfollows:
error
=
1
N
t
N
t

i=1




x
i
− x


+



y
i
− y



.
(12)
(2) Root Mean Square Error. The root m ean square error for
Euclidean distance can be computed as follows:
error
=





1
N
t
N
t

i=1

(
x
i
− x
)
2
+


y
i
− y

2

.
(13)
The root mean square error for Manhattan distance can
be computed as follows:
error
=





1
N
t
N
t

i=1




x
i
− x


+



y
i
− y



.
(14)
(3) Geometric Mean Error. The geometric mean error for
Euclidean distance can be computed as follows:
error
=
N
t





N
t

i=1

(
x
i
− x
)
2
+

y
i
− y

2

.
(15)
The geometric mean error for Manhattan distance can be
computed as follows:
error
=
N
t





N
t

i=1




x
i
− x


+



y
i
− y



.
(16)
Acknowledgments
This work was supported by the National Natural Science
Foundation of China (Grant nos. 61203216, 61273078) and
the Fundamental Research Fund for the Central Universities
of China (N110404030, N110804004, and N110404004).
References
[1] C. Wang and L. Xiao, “Sensor localization in concave environ-
ments,” ACM Transactions on Sensor Networks,vol.4,no.1,
article 3, 2008.
[2]L.M.Kaplan,Q.Le,andP.Molnar,“Maximumlikelihood
methods for bearings-only target localization,” in Proceedings
of IEEE Internat ional Conference on Acoustics, Speech, and
Signal Processing, vol. 5, pp. 3001–3004, Salt Lake City, Utah,
USA, May 2001.
[3] Y. Weng, W. Xiao, and L. Xie, “Total least squares method
for robust source localization in sensor networks using TDOA
measurements,” International Journal of Distributed Sensor
Networks, vol. 2011, Article ID 172902, 8 pages, 2011.
[4] X. Qu and L. Xie, “Source localization by TDOA with r andom
sensor position errors—part I: static sensors,” in Proceedings
of the 15th International Conference on Information Fusion,pp.
48–53, Singapore, July 2012.
[5] X. Qu and L. Xie, “Source localization by TDOA with r andom
sensor position errors—part II: mobile sensors,” in Proceedings
of the 15th International Conference on Information Fusion,pp.
54–59, Singapore, July 2012.
[6] K. C. Ho, “Bias reduction for an explicit solution of source
localization using TDOA,” IEEE Transactions on Signal Process-
ing, vol. 60, no. 5, pp. 2101–2114, 2012.
[7] D. Blatt and A. O. Hero, “Energy-based sensor network source
localization via projection onto convex sets,” IEEE Transactions
on Signal Processing, vol. 54, no. 9, pp. 3614–3619, 2006.
[8] K. Deng and Z. Liu, “Weighted least-squares solutions of
energy-based collaborative source localization using acoustic
array,” International Journal of Computer Science and Network
Securit y, vol. 7, no. 1, pp. 159–165, 2007.
[9] Q. Shi and C. He, “A new incremental optimization algorithm
for ML-based source localization in sensor networks,” IEEE
Signal Processing Letters, vol. 15, pp. 45–48, 2008.
[10] C. Meesookho, U. Mitr a, and S. Narayanan, “On energy-
based acoustic source localization for sensor networks,” IEEE
Transactions on Signal Processing, vol. 56, no. 1, pp. 365–377,
2008.
[11] D. Li and Y. H. Hu, “Energy-based collaborative source local-
ization using acoustic microsensor ar ray,” EURASIP Journal on
Advances in Signal Processing, vol. 2003, article 985029, 2003.
[12] E. Mas¸azade, R. Niu, P. K. Varshney, and M. Keskinoz, “En-
ergy aware iterative source localization for wireless sensor
networks,” IEEE Transactions on Signal Processing, vol. 58, no.
9, pp. 4824–4835, 2010.
[13] M. G. Rabbat, R. D. Nowak, and J. Bucklew, “Robust decen-
tralized source localization via averaging,” in Proceedings of
IEEE International Conference on Acoustics, Speech, and Signal
Processing (ICASSP ’05), vol. 5, pp. V1057–V1060, Philadel-
phia, Pa, USA, March 2005.
[14] D. Ampeliotis and K. Berberidis, “Energy-based model-
independent source localization in wireless sensor networks,”
in Proceedings of the 16th European Signal Processing Confer-
ence, Lausanne, Switzerland, August 2008.
[15] X. Sheng and Y. H. Hu, “Maximum likelihood multiple-source
localization using acoustic energy measurements with wireless
sensor networks,” IEEE Transactions on Signal Processing, vol.
53, no. 1, pp. 44–53, 2005.
[16] W. Meng, W. Xiao, and L. Xie, “An efficient EM algorithm
for energy-based multisource localization in wireless sensor
networks,” IEEE Transactions on Instrumentation and Measure-
ment, vol. 60, no. 3, pp. 1017–1027, 2011.
[17] D. Ampeliotis and K. Berberidis, “Low complexity multiple
acoustic source localization in sensor networks based on
energy measurements,” Signal Processing,vol.90,no.4,pp.
1300–1312, 2010.

10 International Journal of Distributed Sensor Networks
[18] R. Niu and P. Varshney, “Target location estimation in wireless
sensor networks using binary data,” in Proceedings of the 38th
International Conference on Information Sciences and Systems,
pp. 17–19, Princeton, NJ, USA, March 2004.
[19] X. Liu, G. Zhao, and X. Ma, “Target localization and tracking
in noisy binary sensor networks with known spatial topology,”
Wireless Communications and Mobile Computing, vol. 9, no. 8,
pp. 1028–1039, 2009.
[20] M. P. Michaelides and C. G. Panayiotou, “SNAP: fault tolerant
event location estimation in sensor networks using binary
data,” IEEE Transactions on Computers, vol. 58, no. 9, pp. 1185–
1197, 2009.
[21] X. Xu, X. Gao, J. Wan, and N. Xiong, “Trust index based
fault tolerant multiple event localization algorithm for WSNs,”
Sensors, vol. 11, no. 7, pp. 6555–6574, 2011.
[22] K. Lu, X. Xiang, D. Zhang , R. Mao, and Y. Feng, “Localiza-
tion algorithm based on maximum a posteriori in wireless
sensor networks,” International Journal of Distributed Sensor
Networks, vol. 2012, Article ID 260302, 7 pages, 2012.
[23] L. Cheng, C. D. Wu, Y. Z. Zhang, and Y. Wang, “Indoor
robot localization based on wireless sensor networks,” IEEE
Transactions on Consumer Elect ronics, vol. 57, no. 3, pp. 1099–
1104, 2011.
[24] Y. Wang, X. Wang, D. Wang, and D. P. Agrawal, “Range-free
localization using expected hop progress in wireless sensor
networks,” IEEE Transactions on Parallel and Distributed
Systems, vol. 20, no. 10, pp. 1540–1552, 2009.
[25] H. Xu, Y. Tu, W. Xiao, Y. Mao, and T. Shen, “An archimedes
curve-based mobile anchor node localization algorithm in
wireless sensor networks,” in Proceedings of the 8th World
Congress on Intelligent Control and Automation (WCICA ’10),
pp. 6993–6997, Jinan, China, July 2010.
[26] J. Lee, W. Chung, and E. Kim, “Robust DV-Hop algorithm for
localization in wireless sensor network,” in Proceedings of the
International Conference on Control, Automation and Systems,
pp. 2506–2509, Gyeonggi-do, South Korea, October 2010.
[27] J. Zhang, W. Li, D. Cui, X. Sun, and F. Zhou, “Study on
improved DV-Hop node localization algorithm in wireless
sensor network,” in Proceedings of the 5th IEEE Conference on
Industrial Electronics and Applications (ICIEA ’10), pp. 1855–
1858, Taichung, Taiwan, June 2010.
[28] S. W. Lee, D. Y. Lee, and C. W. Lee, “Enhanced DV-Hop
algorithm w ith reduced hop-size error in ad hoc networks,”
IEICE Transactions on Communications,vol.94,no.7,pp.
2130–2132, 2011.
[29] Y. Zheng, L. Wan, Z. Sun, and S. Mei, “A long range DV-
Hop localization algorithm with placement strategy in wireless
sensor networks,” in Proceedings of the 4th International Con-
ference on Wireless Communications, Networking and Mobile
Computing (WiCOM ’08), Dalian, China, October 2008.
[30] J. Lee, W. Chung, and E. Kim, “A new range-free localization
method using quadratic programming,” Computer Communi-
cations, vol. 34, no. 8, pp. 998–1010, 2011.
[31] Z. Zhong and T. He, “RSD: a metric for achieving range-
free localization beyond connectivity,” IEEE Transactions on
Parallel and Distributed Systems, vol. 22, no. 11, pp. 1943–
1951, 2011.
[32] Q. B. He, F. Chen, S. Cai, J. Hao, and Z. Liu, “An effi-
cient range-free localization algorithm for wireless sensor
networks,” Science China Technolog ical Sciences, vol. 54, no. 5,
pp. 1053–1060, 2011.
[33] Y. Chen, W. Chung, and S. Yuan, “Order-based localization
scheme for ad hoc sensor networks,” in Proceedings of the
73rd IEEE Vehicular Technology Conference, pp. 1–5, Budapest,
Hungary, May 2011.
[34] H. Lim and J. C. Hou, “Distributed localization for anisotropic
sensor networks,” ACM Transactions on Sensor Networks, vol.
5, no. 2, article 11, 2009.
[35] P. Brida, J. Machaj, and J. Duha, “A novel optimizing algorithm
for DV based positioning methods in ad hoc networks,”
Elektronika ir Elektrotechnika, no. 1, pp. 33–38, 2010.
[36] O. Chia-Ho, “A localization scheme for wireless sensor
networks using mobile anchors with directional antennas,”
IEEE Sensors Journal, vol. 7, no. 11, pp. 1607–1616, 2011.
[37] H. P. Tan, R. Diamant, and W. K. G. Seah, “A survey of
techniques and challenges in underwater localization,” Ocean
Engineering, vol. 38, no. 14-15, pp. 1663–1676, 2011.
[38] X. Luo, W. J. O’Brien, and C. L. Julien, “Comparative evalu-
ation of Received Signal-Strength Index (RSSI) based indoor
localization techniques for construction jobsites,” Advanced
Engineer ing Informatics, vol. 25, no. 2, pp. 355–363, 2011.
[39] D. J. Suroso, P. Cherntanomwong, P. Sooraksa, and J. Tak-
ada, “Fingerprint-based technique for indoor localization in
wireless sensor networks using Fuzzy C-Means clustering
algorithm,” in Proceedings of the International Symposium on
Intelligent Signal Processing and Communications Systems,pp.
1–5, Chiang Mai, Thailand, December 2011.
[40] S. H. Fang, T. N. Lin, and K.-C. Lee, “A novel algorithm
for multipath fingerprinting in indoor WLAN environments,”
IEEE Transactions on Wireless Communications, vol. 7, no. 9,
pp. 3579–3588, 2008.
[41] N. Swangmuang and P. Krishnamurthy, “An effective location
fingerprint model for wireless indoor localization,” Pervasive
and Mobile Computing, vol. 4, no. 6, pp. 836–850, 2008.
[42] J. Wang, Q. Gao, H. Wang et al., “Differential radio map-based
robust indoor localization,” EURASIP Journal on Wireless
Communications and Networking, vol. 2011, article 17, 2011.
[43] L. Gogolak, S. Pletl, and D. Kukolj, “Indoor fingerprint lo-
calization in WSN environment based on neural network,”
in Proceedings of the 9th IEEE International Symposium on
Intelligent Systems and Informatics, pp. 293–296, Subotica,
Serbia, September 2011.
[44] M. I. Silventoinen and T. Rantalainen, “Mobile station emer-
gency locating in GSM,” in Proceedings of IEEE International
Conference on Personal Wireless Communications, pp. 232–238,
New Delhi, India, February 1996.
[45] M. P. Wylie and J. Holtzman, “The non-line of sight problem
in mobile location estimation,” in Proceedings of the 5th IEEE
International Conference on Universal Personal Communica-
tions Record (ICUPC ’96), pp. 827–831, Cambridge, Mass,
USA, October 1996.
[46] J. Bor ras, P. Hatrack, and N. B. Mandayam, “Decision the-
oretic framework for NLOS identification,” in Proceedings of
the 48th IEEE Vehicular Technology Conference (VTC ’98),pp.
1583–1587, Ottawa, Canada, May 1998.
[47] S. Mazuelas, F. A. Lago, J. Blas et al., “Prior NLOS measure-
ment correction for positioning in cellular wireless networks,”
IEEE Transactions on Vehicular Technology,vol.58,no.5,pp.
2585–2591, 2009.
[48] Y. T. Chan, W. Y. Tsui, H. C. So, and P. C. Ching, “Time-
of-arrival based localization under NLOS conditions,” IEEE
Transactions on Vehicular Technology, vol. 55, no. 1, pp. 17–24,
2006.
[49] S. Venkatraman and J. J. Caffery, “Statistical approach to
non-line-of-sight BS identification,” in Proceedings of the

International Journal of Distributed Sensor Networks 11
5th International Symposium on Wireless Personal Multimedia
Communications, vol. 1, pp. 296–300, Honolulu, Hawaii, USA,
October 2002.
[50] S. Gezici, H. Kobayashi, and H. V. Poor, “Non-parametric
non-line-of-sight identification,” in Proceedings of the 58th
IEEE Vehicular Technology Conference (VTC ’03-Fall), vol. 4,
pp. 2544–2548, Orlando, Fla, USA, October 2003.
[51] K. Yu and Y. J. Guo, “Statistical NLOS identification based
on AOA, TOA, and signal strength,” IEEE Transactions on
Vehicular Technology, vol. 58, no. 1, pp. 274–286, 2009.
[52] L. Cheng, C. D. Wu, Y. Z. Zhang, and Y. Wang, “An indoor
localization strategy for mini-UAV in presence of obstacles,”
International Journal of Advanced Robotic Systems, vol. 2012,
pp. 1–8, 2012.
[53] S. Al-Jazzar, J. Caffery Jr., and H. R. You, “A scattering
model based approach to NLOS mitigation in TOA location
systems,” in Proceedings of the 55th IEEE Vehicular Technology
Conference, pp. 861–865, Birmingham, Ala, USA, May 2002.
[54] L. Liu, P. Deng , and P. Fan, “A TOA reconstruction method
based on ring of scatterers model,” in Proceedings of the 4th
International Conference on Parallel and Distributed Comput-
ing, Applications and Technologies ( PDCAT ’03) , pp. 375–377,
Chengdu, China, August 2003.
[55] W. H. Foy, “Position-location solutions by Taylor-series
estimation,” IEEE Transactions on Aerospace and Electronic
Systems, vol. 12, no. 2, pp. 187–194, 1976.
[56] P. C. Chen, “A non-line-of-sight error mitigation algorithm in
location estimation,” in Proceedings of IEEE Wireless Commu-
nications and Networking Conference, vol. 1, pp. 316–320, New
Orleans, La, USA, September 1999.
[57] L. Cheng, C. D. Wu, Y. Z. Zhang, and H. Chu, “Mobile location
estimation scheme in NLOS environment,” IEICE Electronics
Express, vol. 8, no. 21, pp. 1829–1835, 2011.
[58] S. Venkatraman, J. J Caffery,andH.R.You,“Locationusing
LOS range estimation in NLOS environments,” in Proceedings
of the 55th Vehicular Technology Conference, pp. 856–860,
Birmingham, Ala, USA, May 2002.
[59] I. Povescu, I. Nafomita, P. Constantinou, A. Kanatas, and N.
Moraitis, “Neural networks applications for the prediction of
propagation path loss in urban environments,” in Proceedings
of the 53rd IEEE Semi-Annual Vehicular Technology Conference,
pp. 387–391, Rhodes, Greece, May 2001.
[60] B. L. Le, K. Ahmed, and H. Tsuji, “Mobile location estimator
with NLOS mitigation using Kalman filtering,” in Proceedings
of IEEE Wireless Communications and Networking (WCNC
’03), vol. 3, pp. 1969–1973, New Orleans, La, USA, March
2003.
[61] W. Kim, G. I. Jee, and J. Lee, “Wireless location with NLOS
error mitigation in Korean CDMA system,” in Proceedings of
the 2nd International Conference on 3G Mobile Communication
Technologies, pp. 134–138, London, UK, March 2001.
[62] L. M. Kaplan, “Global node selection for localization in a
distributed sensor network,” IEEE Transactions on Aerospace
and Electronic Systems, vol. 42, no. 1, pp. 113–135, 2006.
[63] L. M. Kaplan, “Local node selection for localization in a dis-
tributed sensor network,” IEEE Transactions on Aerospace and
Electronic Systems, vol. 42, no. 1, pp. 136–146, 2006.
[64] A. Bel, J. L. Vicario, and G. Seco-Granados, “Node selection for
cooperative localization: efficient energy vs. accuracy trade-
off,” in Proceedings of the 5th IEEE International Symposium
on Wireless Pervasive Computing (ISWPC ’10), pp. 307–312,
Modena, Italy, May 2010.
[65] X. J. Yang, K. Y. Xing, K. L. Shi, and Q. Pan, “Dynamic
collaborative algorithm for maneuvering target tracking in
sensor networks,”
Acta Automatica Sinica, vol. 33, no. 10, pp.
1029–1035, 2007.
[66] W. S. Zhang and G. H. Cao, “DCTC: dynamic convoy tree-
based collaboration for target tracking in sensor networks,”
IEEE Transactions on Wireless Communications, vol. 3, no. 5,
pp. 1689–1701, 2004.
[67] X. J. Yang, K. Y. Xing, K. L. Shi, and Q. Pan, “Dynamic
collaborative algorithm for maneuvering target tracking in
sensor networks,” Acta Automatica Sinica, vol. 33, no. 10, pp.
1029–1035, 2007.
[68] Y. E. M. Hamouda and C. Phillips, “Metadata based optimal
sensor selection for multi-target tracking in wireless sensor
networks,” International Journal of Research and Reviews in
Computer Scie nce, vol. 2, no. 1, pp. 189–200, 2011.
[69] A. Bel, J. Lopez Vicario, and G. Seco-Granados, “Real-time
path loss and node selection for cooperative localization in
wireless sensor networks,” in Proceedings of the 21st IEEE
International Symposium on Personal, Indoor and Mobile
Radio Communications Workshops (PIMRC ’10) , pp. 283–288,
Istanbul, Turkey, September 2010.
[70] T. Zhao and A. Nehorai, “Information-driven distributed
maximum likelihood estimation based on Gauss-Newton
method in wireless sensor networks,” IEEE Transactions on
Signal Processing, vol. 55, no. 9, pp. 4669–4682, 2007.
[71] H. Ren and M. Q. Meng, “Power a daptive localization
algorithm for wireless sensor networks using particle filter,”
IEEE Transactions on Vehicular Technology,vol.58,no.5,pp.
2498–2508, 2009.
[72] C. You, Y. Chen, J. Chiang, P. Huang, H. Chu, and S. Lau,
“Sensor-enhanced mobility prediction for energy-efficient
localization,” in Proceedings of the 3rd Annual IEEE Conference
on Sensor and Ad Hoc Communications and Networks, vol. 1,
pp. 565–574, Reston, Va, USA, September 2006.
[73] J. Gribben, A. Boukerche, and R. Pazzi, “Scheduling for scal-
able energy-efficient localization in mobile ad hoc networks,”
in Proceedings of the 7th Annual IEEE Communications Society
Conference on Sensor, Mesh and Ad Hoc Communications and
Networks (SECON ’10), Boston, Mass, USA, June 2010.
[74] Y. Zhu and L. M. Ni, “Probabilistic wakeup: adaptive duty
cycling for energy efficient event detection,” in Proceedings
of the 10th ACM Symposium on Modeling, Analysis, and
Simulation of Wireless and Mobile Systems (MSWiM ’07),pp.
360–367, Chania, Greece, October 2007.
[75] D. K. Goldenberg, P. Bihler, M. Cao et al., “Localization in
sparse networks using sweeps,” in Proceedings of the 12th
Annual International Conference on Mobile Computing and
Networking (MOBICOM ’06), pp. 110–121, Los Angeles, Calif,
USA, September 2006.
[76] X. Ji and H. Zha, “Sensor positioning in wireless ad-hoc sensor
networks using multidimensional scaling,” in Proceedings of
the 23rd Annual Joint Conference of the IEEE Computer and
Communications Societies (INFOCOM ’04), vol. 4, pp. 2652–
2661, Hong Kong, China, March 2004.
[77] Y. Shang and W. Ruml, “Improved MDS-based localization,”
in Proceedings of the 23rd Annual Joint Conference of the IEEE
Computer and Communications Societies (INFOCOM ’04), vol.
4, pp. 2640–2651, Hong Kong, China, March 2004.
[78] K.-F. Ssu, C.-H. Ou, and H. C. Jiau, “Localization with mobile
anchorpointsinwirelesssensornetworks,”IEEE Transactions
on Vehicular Technology, vol. 54, no. 3, pp. 1187–1197, 2005.
[79] M. Erol, L. F. M. Vieira, and M. Gerla, “Localization w ith
Dive’n’Rise (DNR) beacons for underwater acoustic sensor

12 International Journal of Distributed Sensor Networks
networks,” in Proceedings of the 2nd Workshop on Underwater
Networks (WuWNet ’07), pp. 97–100, Montreal, Canada,
September 2007.
[80] U. A. Khan, S. Kar, and J. M. F. Moura, “Distributed sensor
localization in random environments using minimal number
of anchor nodes,” IEEE Transactions on Signal Processing, vol.
57, no. 5, pp. 2000–2016, 2009.
[81] H. Wymeersch, J. Lien, and M. Z. Win, “Cooperative localiza-
tion in wireless networks,” Proceedings of the IEEE, vol. 97, no.
2, pp. 427–450, 2009.
[82] V. Tam, K. Cheng, and K. Lui, “Using micro-genetic algo-
rithms to improve localization in wireless sensor networks,”
Journal of Communications, vol. 1, no. 4, pp. 1–10, 2006.
[83] B. Denis, M. Maman, and L. Ouvry, “On the scheduling of
ranging and distributed positioning updates in cooperative
IR-UWB networks,” in Proceedings of IEEE Internat ional
Conference on Ultra-Wideband (ICUWB ’09), pp. 370–375,
Vancouver, Canada, September 2009.
[84] K. Das and H. Wymeersch, “Censore d cooperative positioning
for dense wireless networks,” in Proceedings of the 21st IEEE
International Symposium on Personal, Indoor and Mobile
Radio Communications Workshops (PIMRC ’10) , pp. 262–266,
Istanbul, Turkey, September 2010.
[85] R. Fujiwara, K. Mizugaki, T. Nakagawa, D. Maeda, and M.
Miyazaki, “TOA/TDOA hybrid relative positioning system
using UWB-IR,” in Proceedings of IEEE Radio and Wireless
Symposium (RWS ’09), pp. 679–682, San Diego, Calif, USA,
January 2009.
[86] M. R. Gholami, S. Gezici, M. Rydstr
¨
om, and E. G. Str
¨
om, “A
distributed positioning algorithm for cooperative active and
passive sensors,” in Proceedings of the 21st IEEE International
Symposium on Personal Indoor and Mobile Radio Commu-
nications (PIMRC ’10), pp. 1713–1718, Instanbul, Turkey,
September 2010.
[87] X. Du, D. Mandala, W. Zhang, C. You, and Y. Xiao, “A
boundary-node based localization scheme for heterogeneous
wireless sensor networks,” in Proceedings of IEEE Mili-
tary Communications Conference (MILCOM ’07), pp. 1–7,
Orlando, Fla, USA, October 2007.
[88] S. Dong, P. Agrawal, and K. Sivalingam, “Localization error
evaluation in heterogeneous sensor networks,” in Proceedings
of IEEE Global Telecommunications Conference, pp. 1–5, New
Orleans, La, USA, December 2008.
[89] C. T. Chiang, P. H. Tseng, and K. T. Feng, “Hybrid unified
kalman tracking algorithms for heterogeneous wireless local-
ization systems,” IEEE Transactions on Vehicular Technology,
vol. 61, no. 2, pp. 702–715, 2012.
[90] C. Shen and D. Pesch, “A heuristic relay positioning algorithm
for heterogeneous wireless networks,” in Proceedings of the
69th IEEE Vehicular Technolog y Conference, pp. 1–5, Barcelona,
Spain, April 2009.
[91] H. Aksu, D. Aksoy, and I. Korpeoglu, “A study of localization
metrics: evaluation of position errors in wireless sensor
networks,” Computer Networks, vol. 55, no. 15, pp. 3562–3577,
2011.

International Journal of
Aerospace
Engineering
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2010
Robotics
Journal of
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Active and Passive
Electronic Components
Control Science
and Engineering
Journal of
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
International Journal of
Rotating
Machinery
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com
Journal of
Engineering
Volume 2014
Submit your manuscripts at
http://www.hindawi.com
VLSI Design
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Shock and Vibration
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Civil Engineering
Advances in
Acoustics and Vibration
Advances in
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Electrical and Computer
Engineering
Journal of
Advances in
OptoElectronics
Hindawi Publishing Corporation
h
ttp://www.hindawi.com
Volume 2014
The Scientic
World Journal
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Sensors
Journal of
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Modelling &
Simulation
in Engineering
Hindawi Publishing Corporation
h
ttp://www.hindawi.com
Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Chemical Engineering
International Journal of
Antennas and
Propagation
International Journal of
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Navigation and
Observation
International Journal of
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Distributed
Sensor Networks
International Journal of