A preview of this full-text is provided by Wiley.
Content available from IET Wireless Sensor Systems
This content is subject to copyright. Terms and conditions apply.
IET Wireless Sensor Systems
Research Article
Hybrid detection algorithm for online faulty
sensors identification in wireless sensor
networks
ISSN 2043-6386
Received on 10th April 2020
Revised 22nd June 2020
Accepted on 10th July 2020
E-First on 3rd August 2020
doi: 10.1049/iet-wss.2020.0053
www.ietdl.org
Walaa Ibrahim Gabr1, Mona A. Ahmed1, Omar M. Salim1
1Department of Electrical Engineering, Benha University, Benha, Egypt
E-mail: omar.salim@bhit.bu.edu.eg
Abstract: Wireless sensor network (WSN) is a developed wireless network consisting of some connected sensor nodes. The
WSN is employed in many fields such as military, industrial, and environmental monitoring applications. These nodes are
equipped with sensors for sensing the environmental variables such as temperature, humidity, wind speed, and so on. In most
applications, WSN is positioned in remote places and harsh environments, where they are most probably exposed to faults.
Hence, faulty sensor identification is one of the most fundamental tasks to be considered in WSN. This study suggests a hybrid
methodology based on mutual information change (MIC) and wavelet transform (WT) for faulty sensor identification. The MIC
method is suggested to study correlation among sensors, while the WT technique is proposed for self-sensor detection. WT is
suitable for analysing non-stationary signals into approximation and detail coefficients. The suggested algorithm performance is
investigated by applying a real case study at an arbitrary location close to Cairo, Egypt. The results of each method are
compared using the true positive rate (TPR), false negative rate, and accuracy measures. Obtained results have shown that
combining MIC and WT techniques can achieve a higher TPR and accuracy reach 100% in most fault types.
1Introduction
The development of sensing abilities and wireless communication
technologies has greatly increased, leading to the development of
inexpensive and low power wireless sensor network (WSN) [1, 2].
WSN is an advanced system composed of a set of sensor nodes.
The major function of each node is to sense, process information,
and communicate this information to their neighbours [1, 3].
The main objective of designing a WSN is to monitor, detect,
and provide useful information about network performance [4].
The architecture of a typical WSN is represented in Fig. 1. Sensor
nodes may be grouped into clusters or cooperatively work together
to complete a common task [3, 5]. In clustering, sensor nodes are
divided into clusters. Each cluster has a leader node called a cluster
head (CH). The CH is responsible for gathering the received data
from the other remaining nodes, which act as cluster member
nodes, hence, CH transmits the aggregated data to the sink node [5,
6]. Each node is equipped with sensor(s) that is/are responsible for
sensing the surrounding environmental variables such as
temperature, wind speed, wind direction, radiation etc. [7, 8]. In
this research work, a study of the relationship among sensors’ data
using mutual information (MI) measure is proposed to identify
faulty sensors in the network.
WSN has gained popularity in recent years [9]. It has major
characteristics over wired networks such as remote sensing, low
cost, rapid deployment, and self-organisation [10]. Due to these
advantages, WSN is used in various applications such as animal
tracking, health monitoring, and industrial applications [3].
Therefore, sensor recordings should be accurate to ensure reliable
operation and avoid faulty data [11]. However, in most
applications, the sensor nodes are deployed in harsh environments
and unstable conditions that could affect the performance of the
WSN [12]. In addition, the sensors are often influenced by noise,
unwanted disturbance, such as electromagnetic interference,
vibration, shot/flicker noise, and environmental noise [13].
Consequently, WSN is vulnerable to faults. A fault is an
undesirable event that results in corrupted data affecting the quality
of the system [14]. Therefore, the identification of faulty sensors in
WSN is necessary [15]. Many fault detection algorithms have been
proposed in the literature to overcome this problem. One of the
suggested solutions to maintain the operation of the network, even
if a fault is present [16], is to estimate missing data of faulty sensor
at a certain area based on the data of other sensors at the correlated
places as explained in [17]. More details about the work related to
fault detection will be described in the literature review section.
The main objective of this paper is to develop a new hybrid
methodology based on the mutual information change (MIC)
method and WT technique to identify faulty sensors. The
advantage of using the WT tool is that every sensor can be
monitored separately. Moreover, it can determine the instantaneous
fault. A comparison is done between both methods to evaluate the
performance of each approach.
This paper is organised as follows: Section 1 introduces the
motivations and defines the problem. Section 2 surveys the
previous work related to fault detection. Section 3 describes an
overview of information theory measures. Section 4 presents a new
technique for signal analysis. Section 5 explains the proposed fault
detection algorithm with a real case study. Section 6 discusses the
simulation results of the selected scenarios, and finally, Section 7
concludes.
Fig. 1 WSN architecture
IET Wirel. Sens. Syst., 2020, Vol. 10 Iss. 6, pp. 265-275
© The Institution of Engineering and Technology 2020
265
