European Journal of Scientific Research

ISSN 1450-216X / 1450-202X Vol. 105 No 2 July, 2013, pp.190-202

http://www.europeanjournalofscientificresearch.com

An Indoor Fingerprinting Localization Approach for

ZigBee Wireless Sensor Networks
Tareq Alhmiedat
Computer Science Department, Zarqa University, Zarqa Street, Jordan
E-mail: [email protected]
Ghassan Samara
Computer Science Department, Zarqa University, Zarqa Street, Jordan
E-mail: [email protected]
Amer O. Abu Salem
Computer Information Systems Department, Zarqa University, Zarqa, Jordan
E-mail: [email protected]
Abstract
Location tracking systems are increasingly becoming the focus of research in the
field of Wireless Sensor Network (WSN). Received Signal Strength (RSS)-based
localization systems are at the forefront of tracking research applications. Radio location
fingerprinting is one of the most promising indoor positioning approaches due to its
powerful in terms of accuracy and cost. However, fingerprinting systems require the
collection of a large number of reference points in the tracking area to achieve reasonable
localization accuracy. In this paper, we propose a fingerprinting localization approach
based on a RSS technique. The proposed system does not require gathering a large number
of reference points and offers good localization accuracy indoors. The implemented
approach is based on dividing the tracking area into subareas and assigning a unique feature
to each subarea through ranging the RSS values from different reference points. In order to
test the proposed system's efficiency, a number of real experiments have been conducted
using Jennic sensor nodes.

Keywords: Wireless Sensor Networks, Tracking, Localization, Fingerprinting, ZigBee.

1. Introduction
Wireless Sensor Network (WSN) has become a vital research area, due to their wide ranging
applications including civilian, industrial, agricultural, and military [1, 2, 3]. A sensor network consists
of sensor nodes which are small in size, low in cost, and have short communication range. Usually, a
sensor device consists of four main subsystems (computing, sensing, communication, and power
supply subsystems).
Researchers have focused on different aspects of WSN, such as hardware design, routing,
security, and localization [4, 5, 6]. One of the critical aspects which needs to be taken into
consideration is localizing mobile targets through distributed sensor networks. Node localization is the
problem of finding the geographical location of a mobile target node (the node with unknown
location), based on other beacon nodes (nodes with fixed known locations).

An Indoor Fingerprinting Localization Approach for ZigBee Wireless Sensor Networks

191

Target tracking applications using WSNs have received much attention recently. This attention
focuses on the need to achieve high localization accuracy, without incurring a large cost, form factor
and power-consumption per node. There are many localization and tracking systems have been
designed and implemented recently, summarized in [7, 8, 9]. A well known localization approach is the
Received Signal Strength (RSS). RSS-based localization systems are one of the most popular and
cheap techniques and are increasingly accepted as a positioning solution for localizing target nodes in
both indoor and outdoor environments. RSS-based localization systems work by converting the Signal
Strength (SS) to a transmitter-receiver using separate distance measurements. However, RSS values
can be affected by walls and obstacles which may reflect and propagate the signals, therefore offering a
non-linear transformation between the RSS values and the location. Due to the aforementioned
limitations, deploying RSS-based localization systems in indoor environments becomes a complicated
task which is difficult to engineer using classical mathematical models.
The RSS information can be used to estimate the distance between the transmitter and the
receiver in two ways: the first uses the signal propagation model to convert SS to a distance
measurement, using previous knowledge about the beacon nodes locations, and deploys a geometry
method to compute the location of target nodes. This is known as a triangulation localization method
[10]. The second approach is based on the behavior of signal propagation and information about the
geometry of the building to convert RSS values into distance values; this is known as a fingerprinting
localization method.
The complexity of applying a triangulation approach arises from the need to accurately obtain
the distance measurements from the RSS as indoor radio signal propagation is very complicated
because of signal attenuation due to distance, the effect of multipath propagation, and penetration
losses through obstacles.
On the other hand, fingerprinting systems require only the collection of RSS values at several
locations to form a database of location fingerprints. Deployment of the fingerprinting-based
localization system is usually divided into two main phases: 1) Offline phase: this includes measuring
the location of a mobile target in several coordinates and storing the collected RSS values at each point
with the corresponding location in a database file; 2) Online phase: the mobile target collects several
RSS values from different beacon nodes in its range and sends the data to a server which applies a
positioning algorithm to estimate the mobile targets location [11].
There are two main challenges to designing and developing a fingerprinting system. First, there
is the problem of collecting the RSS samples and storing them in a database file; this process requires a
long period of time, particularly when the localization system is deployed in a large area. Second, the
searching procedure through the stored samples is time consuming. In this paper, we propose a
fingerprinting-based localization approach which aims to reduce the total number of reference points
that need to be collected in the offline phase while achieving low localization error of between 1 and
3.5 meters. Moreover, the proposed system assigns a unique feature to each small area in order to
reduce the time needed to carry out the searching procedure.
The main contributions of this paper are as follows: a) unlike the existing approaches which
mainly focus on simulation, a new fingerprinting localization approach is introduced practically using
ZigBee sensor nodes; b) it is demonstrated analytically that the proposed localization approach requires
a lower number of reference points to be gathered than the existing neural network based approaches;
c) through experiments, it is shown that the proposed localization method achieves a low level of
localization error in complex environments.
This paper is organized as follows: In Section 2, the related systems proposed recently are
introduced. In Section 3, a new fingerprinting localization approach is presented, while Section 4
reviews the implementation of the proposed system on real ZigBee-based sensor devices. Section 5
offers the evaluation of the proposed fingerprinting localization system and a discussion is presented in
Section 6. Finally, Section 7 draws conclusions regarding the proposed system and presents
suggestions for future work.

192

Tareq Alhmiedat, Ghassan Samara and Amer O. Abu Salem

2. Related Work
Even though WSNs were not designed and developed for the purpose of localization applications,
measuring the RSS values for each transmitted message could offer localization information of mobile
targets. Fingerprinting localization systems are relatively simple and cheap compared to other methods
such as angle-of-arrival (AOA), time-of-arrival (TOA), and time difference of arrival (TDOA) systems,
as these systems require attaching additional sensor device to each beacon node. In this section, the
fingerprinting-based localization approaches are categorized into two categories based on the
localization method used in the estimation phase: database-based and neural network-based methods.
First, the database-based approaches are considered. The systems proposed in [12, 13] include a
WLAN-based indoor localization system based on fingerprinting method. The developed approaches
are based on collecting RSS values from several reference points distributed over the tracking area and
storing them in a database file. The mobile target estimates its location by comparing the RSS values
collected from the beacon nodes and the RSS values stored in the database.
RADAR is a well-known tracking system proposed in [14], which operates by recording and
processing information regarding the signal strength at multiple base stations in order to provide
overlapping coverage of the area of interest. The work presented in [15] includes a research and a
development of a fingerprinting-indoor positioning system using the received signal strength, which its
based on two different protocol stacks: BitCloud and OpenMac.
Cortina is a distributed real time location system designed to track people indoors based on
IEEE802.15.4 radio standard and RFID technology. Cortina avoids the need for manual calibration by
adopting a collaborative approach that uses RSS measurements collected by the fixed nodes [16]. A
hybrid localization approach using WLAN, which consists of two main stages: first, a fingerprinting
method, with a fast training phase in order to obtain the location for the mobile target, indicates which
room the mobile target is located in. Second, a triangulation is used to compute the mobile target
location precisely. It was shown that the proposed hybrid system is more accurate than the
triangulation approach. However, the proposed hybrid system offers lower localization accuracy than
the fingerprinting method [17]. In [18], a hybrid localization approach based on WSN was proposed,
that integrates two techniques: RF mapping and cooperative ranging to enhance the localization
accuracy.
In order to reduce the time required for the searching process in the online phase. The work
proposed in [19] includes a new search strategy for radio fingerprint matching method, which
significantly reduces the search operations with a little effect on the localization accuracy.
The second approach is neural network-based. The work presented in [20] includes an indoor
localization system based on a modular multi-layer perceptron, where three neural network modules
were designed in order to cover the absence of signals from access points. The proposed work in [21]
includes a fingerprinting-based localization system with a localization error of 3 meters, where a total
number of 5 reference points were needed for the training process in the offline phase. In [22], the
authors compared the performance of three different families of neural networks for localization
applications using WSN: Multi-Layer Perceptron (MLP), Radial Basis Function (RBF), and Recurrent
Neural Networks (RNN). Through experiments, RBF module was the best choice as it offered high
localization accuracy, and minimal computational and memory requirements.

3. Radio Frequency Fingerprinting Approach

The indoor fingerprinting positioning approach developed in this paper consists of three main phases:
the Creation of the fingerprint table phase, the Feature identification phase, and the Estimation phase.
The first two phases are carried out during the offline stage while the third one is carried out in the
online stage. Figure 1 depicts the proposed approach and Table 1 includes the definitions of several
parameters.

An Indoor Fingerprinting Localization Approach for ZigBee Wireless Sensor Networks

193

Figure 1: RF Fingerprinting approach

}
Table 1:

A Sample of Companies Listed in Tehran Stock Exchange

Name
Beacon node
Reference point
Estimated point
Criteria

Parameter
b
r
e
CAi

Definition
A node with fixed position and known coordinates (x, y)
A point with known RSS values collected previously from beacon
nodes, and identified (x, y) coordinates
A point which needs to be found based on the RSS values
collected from beacon nodes
A set of features distinguishes subarea Ai from other subareas

3.1. Creation of Fingerprint Table Phase

Assume the tracking area is divided into grid points, and the coordinate for each grid point is Pi = ( xi , yi ) ,
where n is the total number of grid points. Also m is the total number of reference points,
where m < n . This phase includes two sub-stages:
1. Collect RSS values from several beacon nodes {b1 , b2 ,, bz } at each reference point rk and
store them in a database file. The collected reference points have to be spread as evenly as
possible in the tracking area of interest.
2. Divide the tracking area of interest into subareas as { A1 , A2 ,, Au } , where u is the total number
of subareas.
In the first stage, the mobile target t goes through a total number of reference grid points m . The
mobile target t starts to receive location messages from beacon nodes at each reference point
rk = ( xk , yk ) . Let rss bkj refers to the average RSS values from the bj th beacon node at rk reference
i =1,2,..., n

point. We can establish a RSS vector at each reference grid point as Sk =

k
k
k
k
rssb , rssb , rssb ,..., rssb
z
1
2
3

} , where

is the total number of beacon nodes.

The second stage involves dividing the tracking area into subareas based on assigning a unique
feature to each subarea. There are three reasons behind dividing the tracking area into subareas. First, it
sometimes enhances the localization accuracy, by searching the reference points which set in the same

194

Tareq Alhmiedat, Ghassan Samara and Amer O. Abu Salem

subarea where the mobile target settles in. Second, any changes in the environment after the collection
phase can be recovered by recollecting reference points from that subarea. And third, dividing the
search area into smaller subareas drastically reduces the search process and space. Based on the
previous location of a mobile target, only the last known subarea and the immediately surrounding
reference points need to be searched to find the new location of the mobile target.
A labeled training set (including input signals, corresponding output locations, subareas and
features) is stored in the database, as shown in Table 2.
Table 2:

Structure of database

Vector

RSS values

Corresponding
subarea

Feature

Location

A1

CA1

(x1, y1)

A2

CA2

(x2, y2)

Al

CAl

(xs, ys)

{rss1b1 , rss1b2 , rss1b3 ,..., rss1bz }

{
}
{rssbn1 , rssbn2 , rssbn3 ,..., rssbnz }

S1

rss 2 , rss 2 , rss 2 ,..., rss 2

b1
b2
b3
bz

S2
Sn

3.2. Feature Identification Phase

In this phase, a unique feature is identified for each single subarea. The features can be identified based
on the RSS values collected from adjacent beacon nodes {b1 , b2 ,, bz } at the reference points in the
subarea Ai .Assume that the set

represents the beacon nodes, such that B = {b1 , b2 ,, bz } , and

} , where

set of categories for the subarea A such that

y
1
2
C A = Rab , Rab ,..., Rab
z
i
1
2

values for beacon node bk in the subarea

Let rss Ai function as follows:

Ai .

i
Rab
k

C A is
i

is the range of the RSS

(1)

rss A : R C A
i
i

However, there are N different subareas and consequently there are N different feature
functions. For instance, assume that R = {r1 , r2 ,, bs } represents reference points in a given subarea Ai .
There is a set of features
where C A

= Rab , Rab ,..., Rab

z
1
2

the reference points

R in

C A that
i

} , and

a subarea

Rab
j

A
i.

distinguishes the subarea

A
i

from other subareas,

is the range of RSS values between a beacon node

For instance,

Ra

bi

bj

and all

( 62 75) .

3.3. Estimation Phase

In this phase, the mobile targets location is estimated. This phase includes two main sub-stages:
1. Identify the subarea Ai where the mobile target is located in, based on comparing the
collected RSS values with the identified features assigned to each subarea,
2. Find out the nearest three reference points to the estimated point e in the identified subarea
A based on the difference in the RSS readings in the selected subarea.
i
In the first stage, to locate the position of a mobile target, the RSS value is recorded between
the mobile target t and beacon nodes {b1 , b2 ,, bz } . in the mobile targets communication range. The
RSS values obtained for

are compared to the RSS values for known positions (reference points)

An Indoor Fingerprinting Localization Approach for ZigBee Wireless Sensor Networks

195

stored in the database file. The subarea with the closest RSS values is assumed to be where t is
located.
In the second stage, to be more precise, to locate the mobile target in the subarea, the RSS
values of the mobile node are compared with the features of the beacon nodes within the subarea based
on Equation 2. The Three Nearest Neighbors based on Feature identification function (3NNF) is used
to triangulate the mobile targets position.

e
r
diff rss , rss =
i
j

w
e
r
rss rss
k
k
k =1

(2)

4. Experiments
In order to evaluate the proposed fingerprinting localization approach, several experiments were
conducted on real sensor nodes. In this section, the main features of the proposed system including the
Graphical User Interface (GUI), mobile and beacon nodes modules, and the experimental test-beds,
are illustrated.
4.1. Graphical User Interface (GUI)
In order to facilitate the collection and segmentation processes, the map of the tracking area of interest
is stored on a laptop and a user interface based on Visual Basic.Net, was designed. In this section, the
collection and segmentations processes are discussed in details.
4.1.1. Collection Process
In the first phase, the user has to distribute the beacon nodes over the tracking area of interest, and then
collect the RSS readings at several reference points. This can be achieved by a single click on the
displayed map which selects the reference point on the map and then stores the collected RSS values
from each beacon node. A total number of 70 and 60 measurement points are identified and collected
from distinct physical locations on test-beds 1 and 2 respectively. The graphical user interface of the
proposed system is shown in Figure 2.
Figure 2: Graphical user interface for the proposed system

196

Tareq Alhmiedat, Ghassan Samara and Amer O. Abu Salem

4.1.2. Segmentation Process

Dividing the tracking area into subareas can be achieved in one of two ways: manual or autonomous.
The manual selection can be processed by selecting the desired subarea on the displayed map. The
system is designed to check whether there is a common feature among reference points, whereas, in the
autonomous process, the system divides the tracking area into subareas randomly and then checks the
RSS values collected from reference points in that subarea. If these reference points have a common
feature, then the system will consider that area as a single subarea, and separate it from the other
subareas. Otherwise, the selected subarea will be reduced and the previous step is repeated till all
subareas are considered.
In the estimation phase, the mobile target collects RSS values from beacon nodes and transfers
these values to the sink node; the 3NNF algorithm is applied at the server side in order to display the
final position on a map.
4.2. Mobile and Reference Node Modules
In our experiments, we used a JN5139-EK010 sensor node platform (depicted in Figure 3) for both the
mobile target and beacon nodes. This module offers low power-consumption, low processor overheads,
and a low cost platform for WSNs.
Figure 3: JN5139 Jennic sensor module

The proposed system was implemented through ZigBee network standard. ZigBee standard
offers 3 main roles: coordinator, router and end-device nodes. In our experiments, the mobile target
and beacon nodes were considered as router devices, while the sink node implemented the coordinator
of the ZigBee network, which was responsible for collecting localization information from mobile
target and transferring it to a laptop via a serial cable.
4.3. Experimental Test-beds
The proposed approach was tested in two different experimental test-beds in order to test its efficiency
and accuracy. The experimental Test-bed 1 was located in the 1st floor of the Holywell Park building at
Loughborough University. Its layout has dimensions of (41.5 11.3 m), as shown in Figure 4. Testbed 1 includes obstacles and walls, which might affect the localization accuracy. The experimental
Test-bed 2, depicted in Figure 5, is located in the Sir David Wallace Sports Hall at Loughborough
University; it has dimensions of (30.5 11.3 m). For both test-beds, the origin of the coordinate
system (0, 0) was placed at the left bottom corner.

An Indoor Fingerprinting Localization Approach for ZigBee Wireless Sensor Networks

197

Figure 4: Test-bed 1-FK office test environment

Figure 5: Test-bed 2-Sports hall test environment

5. Testing Results and Performance Evaluation

To verify the validity of the proposed system described in this paper, the system was implemented
through real experiments conducted in indoor environments using Jennic sensor nodes. The proposed
system was evaluated through measuring the localization error, the effect of the collected number of
reference points on the localization accuracy, and the efficiency of the proposed segmentation process.
5.1. Localization Accuracy
Localization accuracy can be described as the distance between the real and estimated locations. The
localization accuracy was evaluated based on two factors: First, the effect of collected samples' number
on the localization accuracy, since it was evaluated based on two different methods (3NNF localization
method and the RBF neural network module). According to [22], the RBF module offers the greatest
efficiency in terms of localization accuracy and memory requirements.
As shown in Figure 6, the localization accuracy for the approach proposed in this paper (3NNF)
is better than the RBF neural network method as the RBF method needs to be trained at every single
grid point; and this requires gathering a large number of reference points through the offline phase.
Whereas the 3NNF method estimates the mobile targets position in two stages: it first finds out the
subarea where the mobile target is located in, and then triangulates the target's location using the
nearest three neighbor nodes.

198

Tareq Alhmiedat, Ghassan Samara and Amer O. Abu Salem

Figure 6: The localization error using 3NNF method in Test-bed 1
12
r

10

y-axis

8
6
4
2
0
0

10

15

20

x-axis

25

30

35

40

45

Second, the localization accuracy is evaluated based on the localization method used in the
estimation phase. Figure 6 shows the localization error for the proposed fingerprinting localization
method when 3NNF was used. The localization error was between 1 and 3 meters. However, the
localization error was between 1.5 and 4 meters when the RBF method was used in the estimation
phase, as depicted in Figure 7. In both figures, the number of reference points ( r ), test points ( t ), and
estimated points ( e ) are depicted.
Figure 7: The localization error using the RBF method in Test-bed 1
12

10

y-axis

6
4
2

0
0

10

15

20x-axis 25

30

35

40

45

5.2. Testing the Efficiency of the Segmentation Process

Assigning unique features to each subarea is a critical task. As previously discussed, the segmentation
process can be achieved in two ways (manual or autonomous). In this section, the ability of our
proposed system to divide the whole tracking area into subareas is evaluated. The proposed system was
tested in two different test-beds.
The autonomous segmentation process works efficiently in open environments, as in Test-bed
1. Figure 8 depicts the RSS values from each beacon node in each subarea. However, deploying an
autonomous process in complex environments where obstacles and walls are situated is a challenging
task as the RSS behaves irregularly in different areas. Thus, it was a challenging task for the
autonomous segmentation to achieve the division process in complex environments. The proposed
autonomous process was not able to divide Test-bed 2 into subareas; a manual segmentation was used
instead. Figure 9 shows the irregular distribution of RSS values in a complex environment. Both
figures show how the tracking area is divided into subareas.

An Indoor Fingerprinting Localization Approach for ZigBee Wireless Sensor Networks

199

Figure 8: Testing the efficiency of the autonomous segmentation process in Test-bed 2

rn
2

rn
1

rn
3

Figure 9: Testing the efficiency of the autonomous segmentation process in Test-bed 1

rn
1

rn
2

rn
3

6. Discussion
Obliviously, the need for more reference points and measurements indicates that the offline stage is a
critical task in terms of time and labor. Neural networks have been used in several fingerprinting
localization approaches [20, 21, 22] in order to reduce the total number of the collected reference
points, and therefore trim down the time needed to complete the offline phase. Conversely, through
experiments, it has been shown that neural network-based approaches (such as RBF) performs well in
the area on which it has been trained, and consequently, neural network modules need to be trained at
each grid point in order to offer reasonable localization accuracy. This procedure adds time and power
consumption to the whole WSN.
The approach presented in this paper does not require the collection of a large number of
reference points as it's based on identifying a unique feature for each subarea. Consequently, the
searching procedure can be processed in a short time period compared to [20, 21, 22].
An efficient WSN-based fingerprinting localization approach was implemented in [22]. The
implementation of this approach was limited in scale as it was implemented in a small area (300 300
cm). Conversely, the approach proposed in this paper was deployed in two applicable areas (Test-bed 1
and Test-bed 2) with sizes of (41.5 11.3 m) and (30.5 11.3 m) respectively.
A significant localization system was proposed in [14] which offers good localization accuracy.
However, it is based on the nearest neighbor method in order to compute the targets coordinates. This
includes finding out the two nearest points to the estimated reference point based on the collected RSS
values. However, the 3NNF first identifies the subarea where the target node is located in, and then
triangulate its location using the nearest neighbor reference points. Figure 10 shows the localization
error for both the 2 nearest neighbors and 3NNF methods.

200

Tareq Alhmiedat, Ghassan Samara and Amer O. Abu Salem

Figure 10: The distance error based on the number of reference points (r)

The proposed system includes dividing the whole area into subareas based on unique features
which uniquely distinguish each subarea. For most previous fingerprinting systems, any changes to the
environments, which alter the collected and stored features at the scenes, require recollecting the
predefined data. However, the approach proposed in this paper does not require reference points to be
recollected from the whole tracking area in the case of changes. It only requires data to be recollected
from the subarea which has been changed. Table 3 compares the existing fingerprinting localization
systems.
Table 3:

Evaluating the existing fingerprinting localization systems

Localization method
Approach [12]
Approach [13]
RADAR [14]
Approach [17]
Approach [18]
Approach [19]
Approach [22]
RBF approach
3NNF approach

Area size (meter)

Hall
43.5 22.5
23 17.5
-

25.5
0.3
41.5
41.5

24.5
0.3
11.3
11.3

Number of collected
reference points
132
441
70
>100
300
200
121
70
70

Localization error (meter)

14
0.5 7.8
23
1.78
0.2 2.2
23
0.1 0.4
16
1 3.5

7. Conclusion and Future Work

RSS-based localization systems are more competitive in terms of both accuracy and cost compared to
other localization systems. Fingerprinting methods suffer from the fact that they require a large
database and a long training phase; this increases the size and computational burden of the database. In
this paper, three advantages were gained over the existing approaches. First, the total number of
reference points that had to be collected during the offline phase was reduced. Second, the proposed
approach offers good localization accuracy (1-3 meters). Third, a problem which might arise when any
changes on the tracking area has been overcome by dividing the tracking area into subareas.
In this paper, the researchers have designed and implemented a segmentation process in order
to divide a tracking area into small areas. The proposed autonomous method only works well in open
environments and cannot be used in complicated ones. Manual segmentation can be used in any
environment but it requires additional effort and time, in addition to experience regarding the system
and how to divide the tracking area. One of the most important developments and improvements that

An Indoor Fingerprinting Localization Approach for ZigBee Wireless Sensor Networks

201

should be added to research in the future is to improve the autonomous segmentation process by
adopting Artificial Intelligence, in addition to increasing localization accuracy by adopting other
localization methods.

References
[1]

[2]

[3]

[4]

[5]

[6]

[7]
[8]
[9]

[10]
[11]

[12]
[13]

[14]

[15]
[16]

Hamdi, M., Boudriga, N., and Obaidat, M. 2008. WHOMoVeS: An optimized broadband
sensor network for military vehicle tracking, International Journal of Communication
Systems, Vol. 21, Issue 3, pp. 277-300, ISSN:1074-5351.
Son, B., Her, Y., and Kim, J. 2006. A design and implementation of forest-fires surveillance
system based on wireless sensor networks for South Korea Mountains. Int. J. Comput. Sci.
Netw. Secur.(IJCSNS) 6, 124130.
Antoine-Santoni, T., Santucci, J., Gentili, E., Silvani, X. and Morandini, F. 2009. Performance
of a protected wireless sensor network in a fire. Analysis of fire spread and data transmission.
Sens. J. 9, 5878-5893.
Alhmiedat, T. A. and Yang, S., 2009. Tracking Multiple Mobile Targets Based on ZigBee
Standard, Proceedings of the 35th Annual Conference of the IEEE Industrial Electronics
Society, Porto, Portugal, Nov. 2009.
Abutaleb, A., Pradhan, D. and Kocak. T. 2009. A Technique to Identify and Substitute Faulty
Nodes in Wireless Sensor Networks. Sensor Technologies and Applications, 2009.
SENSORCOMM'09. Third International Conference on. IEEE.
Abusaimeh, H. and Yang, S. 2009, Reducing the transmission and reception powers in the
AODV. Networking, Sensing and Control, 2009. ICNSC'09. International Conference on.
IEEE.
Alhmiedat, T. Yang, S. 2007, A Survey: Localization and Tracking Mobile Targets through
Wireless Sensor Network, Proceedings of the PGNet conference, Liverpool, UK, pp. 48, June.
Amitangshu, P. 2010, Localization algorithms in wireless sensor networks: Current
approaches and future challenges. Network Protocols and Algorithms 2.1. 45-73.
Alhmiedat, T. A., Abutaleb, A. and Bsoul, M. 2012, A Study on Threats Detection and
Tracking for Military Applications Using WSNs, International Journal of Computer
Applications, 40(15):12-18.
Alhmiedat, T. A. and Yang. S. 2008. A ZigBee-based Mobile Tracking System Through
Wireless Sensor Networks, Int. Advanced Mechatronic Systems, 1(1), pp. 63-70.
Hu, L. and Evans. D. 2004, Localization for mobile sensor networks, Proceedings of the
Tenth Annual International Conference on Mobile Computing and Networking (MOBICOM),
Philadelphia, PA. Sept.
Li, B., Salter, J., Dempster, A. and Rizos, C., 2006, Indoor Positioning Techniques Based on
Wireless LAN, Proceedings of AusWireless, Sydney, Australia..
Small, J., Smaliagic, A. and Siewiorek, D. 2000 Determining user Locations For Context
Aware Computing Through the Use of a Wireless LAN Infrastructure, Institute for Complex
Engineered Systems, Carnegie Mellon University, Pittsburgh, PA 15213, USA.
Bahl, P. and Padmanabhan, V. 2000. Radar: an in-building RF-based user location and
tracking system," in Proceedings of 19th Annual Joint Conference of the IEEE Computer and
Communications Societies (INFOCOM '00), vol. 2, pp. 775-784, Tel Aviv, Palestine.
Martin, J., Rodrguez, A., Zubiete, E., Romero, O. and Guilln, S. 2013. Fingerprint Indoor
Position System Based. Journal of Networks, 8(1), 37-44.
Sun, Z., Farley, R., Kaleas, T., Ellis, J., and Chikkappa, K. 2011. Cortina: Collaborative
context-aware indoor positioning employing rss and rtof techniques, In Pervasive Computing
and Communications Workshops (PERCOM Workshops), 2011 IEEE International Conference
on (pp. 340-343). IEEE.

202

Tareq Alhmiedat, Ghassan Samara and Amer O. Abu Salem

[17]

Li, B., Dempster, A., Rizos, C. and Barnes, J. 2005. Hybrid Method for Localization using
WLAN, Spatial Sciences Conference, Melbourne, Australia.
Seet, B., Zhang, Q., Foh, C and. Fong, A. 2009. Hybrid RF mapping and ranging based
localization for wireless sensor networks, in Sensors 09: The 8th Annual IEEE Conference on
Sensors.
Li, H., Zhao, X. and Tan. M. 2011. A New Search Strategy of Radio Fingerprint Matching
Method in Wireless Sensor Network, First International conference on Robot, Vision and
Signal Processing.
Ahmad, U. et al., 2006. In-building Localization using Neural Networks, IEEE International
Conference of Intelligent Systems, Islamabad, Pakistan.
Battiti, R., Nhat, T. and Villani, A. 2002. Location-aware computing: a neural network model
for determining location in wireless LANs, Tech. Rep. DIT-02-0083, University of Trento,
Trento, Italy.
Shareef, A., Zhu, Y. and Musavi, M. 2008. Localization Using Neural Networks in Wireless
Sensor Networks, 1st International Conference on Mobile Wireless MiddleWARE, Operating
Systems, and Applications, Innsbruck, Australia.

[18]

[19]

[20]
[21]

[22]