3DPHDV-Hop: A New 3D Positioning Algorithm
using Partial HopSize in WSN
Rui-Jin Wang , Zhi-Guang Qin , Ja-Hao Wang , Yang Zhao , Wei-Wei Ni
School of Computer Science and Engineering, University of Electronic Science and Technology of China
Chengdu 610054, China
Email: wrj8882003@163.com
Abstract—This paper presents a new three-dimensional posi-
tioning algorithm, named 3DPHDV-Hop, which can significantly
improve the performance in terms of positioning error and
coverage by using self-tune selection of projection plane and
exploiting partial average hop size instead of global average hop
size in wireless sensor network. The simulation results show that
the accuracy of our proposed algorithm can notably enhance the
positioning coverage, compared to APIS algorithm at a maximum
percentage of up to 45.2%, ‘Simple 3D DV-Hop’ at a percentage
of 35.4%, and ‘3D-MDS’ at a percentage of 32.3%. Therefore,
our algorithm has the larger potential which can be used in much
more fields under the real deployment environment of WSN.
I. INTRODUCTION
Wireless sensor networks (WSNs) consist of hundreds and
sometimes thousands of inexpensive, low-power and tiny sen-
sors. These sensors usually have very limited sensing, commu-
nicational and computational capabilities [1], even if they can
organize themselves and cooperate to achieve a given task.
In recent years, worldwide attention of WSN was attracted
[2]. WSNs have been widely used in many applications such
as military target tracking [3], natural disaster relief [4], haz-
ardous environment exploration [5], etc. It is very important
to accurately positioning target (i.e., unknown nodes) in these
applications. To address the localization problem, many local-
ization algorithms have been proposed [2, 6–9]. The sensors in
WSNs are usually deployed in mountains, forests, oceans and
air. Therefore, it is characteristic of these sensor nodes to be
distributed in three dimensional space. However, the majority
of the current localization techniques are designed assuming
2D deployments, which obviously cannot meet the needs of
the practical application. As a result, it is very essential to
locate target nodes in three-dimensional (3D) space.
This paper presents an improved three-dimensional posi-
tioning algorithm, called 3DPHDV-Hop, which is based on
self-tune selection of projection plane and partial average hop
size instead of global average hop size. The simulations results
prove that 3DPHDV-Hop algorithm can significantly improve
the positioning coverage compared to ‘Simple 3D DV-Hop’,
‘APIS’, and ‘3D-MDS’. At the same time, the proposed algo-
rithm can enhance positioning accuracy considerably without
extra overheads in term of computation and communication.
The remainder of this paper is organized as follows. Related
work is presented in Section II, followed by essential back-
ground on 3D DV-Hop positioning algorithm in Section III,
followed by the design issues about 3DPHDV-Hop algorithm
in Section IV. Experimental results and performance analysis
are presented in Section V. The conclusion and future work
are presented in Section VI.
II. R
ELATED WORK
Some research has been conducted on three-dimensional
positioning algorithms of WSN under different assumptions
and in different environments. For example, AbdelSalam et
al. presented a 3D-localization technique [1], coming with
a terrain modeling capability which is based on Delaunay
triangulation. In this technique, the horizontal plane first is
determined to contain the sensors, including unknown nodes
and anchor nodes. Then, the nearest three anchor nodes are
projected onto the horizontal plane determined in the first step.
Finally, RSSI-based distance measurements and trilateration
techniques are used to fully localize the unknown nodes.
However, only horizontal plane is selected as the one that
contains sensors, which is not reasonable in practice. Lu et
al. proposed a three dimensional locate scheme [10], named
APIS, which is based on sphere intersections in WSN. In
APIS, the anchor nodes are regarded the as the center of
a sphere and its radius is the distance between the current
anchor node and the other anchor nodes. But the coverage and
positioning accuracy of this algorithm are influenced seriously
by the distribution of anchor nodes. Liang et al. presented a
limited space within the three-dimensional positioning algo-
rithm [11], in which all the anchor nodes are on the bottom
plane and the procedure of localization for the whole network
is from the bottom to the top. Furthermore, the procedure
can be done in all directions from the area where anchor
nodes are clustered. Another three-dimensional positioning
algorithm, called 3D-MDS, was proposed in [12], which is
based on multidimensional scaling. It combines the experience
attenuation model of the received 3D signal strength (RSS) and
the shortest path method, and decomposes dissimilarity matrix
by a lightweight matrix decomposition method. However, the
cost of this algorithm in terms of computation and communi-
cation is very expensive and the configuration of hardware for
implementing the algorithm is very high.
In these three-dimensional positioning algorithms intro-
duced above, all of them are based on the global average jump
distance to locate the target nodes. With the increase of hops,
91