Abstract
Localization is an essential and important research issue in wireless sensor networks (WSNs). Most localization schemes
focus on static sensor networks. However, mobile sensors are required in some applications such that the sensed area can be
enlarged. As such, a localization scheme designed for mobile sensor networks is necessary. In this paper, we propose a localization
scheme to improve the localization accuracy of previous work. In this proposed scheme, the normal nodes without location information
can estimate their own locations by gathering the positions of location-aware nodes (anchor nodes) and the one-hop normal nodes
whose locations are estimated from the anchor nodes. In addition, we propose a scheme that predicts the moving direction of sensor
nodes to increase localization accuracy. Simulation results show that the localization error in our proposed scheme is lower than the
previous schemes in various mobility models and moving speeds.
1 INTRODUCTION
IN recent years, wireless sensor networks (WSNs) [1], [5]
have been widely used in a wide range of applications
such as military operations, medical treatments, and the
monitoring of animal activity and the environment in the
forest. The basic assumption in many applications is that
sensor nodes have to know their positions. For example,
the sensed data must combine with location information,
for a server instantly to know where an event has
happened. In order to get sensors’ positions, one simple
and precise solution is that each sensor node must carry
Global Positioning System (GPS) equipment. Unfortunately,
it is too expensive to realize and is useless indoors.
Moreover, most applications require coarse localization
accuracy. As such, the reasonable solution is that some
nodes of sensor network should be equipped with a GPS
device, while the others get their positions automatically
by a localization scheme. In general, the location-aware
nodes are called anchor nodes, and the remaining nodes are
called normal nodes.
Many localization schemes have been proposed in the
past few years. Most of them are designed for static sensor
networks [11], [13], [14], [20], [26]. However, some
applications assume that sensors are mobile and locationaware.
For example, in target tracking, the sensor nodes
know their areas by tracking locations of moving objects.
In addition, sensor nodes are mobile for enlarging the
sensing region. Thus, a designed localization scheme for
mobile sensor networks is necessary. A Monte Carlo
Localization (MCL) scheme specifically designed for a
mobile sensor network is proposed in [12]. In MCL, all
sensor nodes are mobile. Each normal node collects the
locations of its one-hop and two-hop anchor nodes via
message exchange, and constructs a new possible location
set in each time slot. The possible location set consists of
various coordinates where the normal node may locate.
The possible locations are also constrained by the communication
range of anchor nodes and the moving region of
location set in the previous time slot. However, the
localization error with low anchor density in MCL does
not work well. The Mobile and Static sensor network
Localization (MSL) [19] is one another range-free algorithm
that uses the Monte Carlo method. MSL improves
localization accuracy by using the location estimation of all
neighbors (not just anchor nodes). The above methods are
time-consuming because they need to keep sampling and
filtering until enough samples are obtained to construct a
new possible location set in each time slot. A boundingbox
(BB) method used to reduce the scope of searching the
candidate samples is proposed

Download full report
http://googleurl?sa=t&source=web&cd=3&ve...210109.pdf%3Farnumber%3D5210109&ei=QjMxTo_gA5LKiALJ9vS0Bg&usg=AFQjCNHSjMh0m_m0_0vHZY5KWr-C72xNSg&sig2=9uGn9nPGFazVOQnVXjX6ug