See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/263684916

Wireless Sensor Networks: A Review

Article in International Journal of Sensors Wireless Communications and Control · December 2013
DOI: 10.2174/22103279112029990001

CITATIONS READS
30 5,062

2 authors, including:

C.P. Gupta
Rajasthan Technical University
42 PUBLICATIONS 244 CITATIONS

SEE PROFILE

All content following this page was uploaded by C.P. Gupta on 20 August 2019.

The user has requested enhancement of the downloaded file.

 Send Orders for Reprints to reprints@benthamscience.net
International Journal of Sensors, Wireless Communications and Control, 2013, 3, 25-36 25

Wireless Sensor Networks: A Review

C.P. Gupta1 and Arun Kumar2*

1
Department of Computer Engineering, Rajasthan Technical University, Kota, India; 2Department of Mathematics,
Government College, Kota, India

Abstract: Wireless Sensor Networks (WSNs) are being increasingly deployed in military, health care, health monitoring,
environmental and several other applications. However, design and deployment of WSNs have several challenges. In this
paper, we present the challenges in deploying WSNs for different applications. The challenges in the design of communi-
cation infrastructure have also been discussed. Effects of using mobile base stations have been discussed for improving
the lifetime of the network.
Keywords: Base Station, Fault Tolerance, Routing, Wireless Sensor Networks.

1. INTRODUCTION further processing and decisions. WSNs have great potential

for many applications such as habitat monitoring, intrusion
In the recent years, Sensor networks have been deployed detection and target tracking and surveillance, oceanography,
for a variety of applications. In few cases viz. fire safety in
environmental monitoring, structural health monitoring, in-
high rise buildings, the sensors are connected through wires.
frastructure monitoring, precision agriculture, biomedical
However, wired networks are rare and do not find large ap-
health monitoring, and hazardous environment exploration
plications. In most cases, the deployment prescribes for a
and seismic sensing. Sensors have been thought to be ex-
wireless environment. As a consequence, the sensor net-
tremely useful to monitor health of patients. Wearable sen-
works use wireless communication to send the collected data sors have gained interest and found applications in elder
and are thus termed as Wireless Sensor Networks (WSN).
care, emergency response, studying athletic performance,
Formally, a wireless sensor network (WSN) in its simplest
gait analysis, and activity classification. WSN have been
form can be defined as a network of (possibly low-size and
deployed for the systems of vessels and also for traffic con-
low-complex) devices denoted as nodes that can sense the
trol on roads [1, 2]. Deployment of WSN for monitoring
environment and communicate the information gathered
infrastructure like bridges, dams, buildings, pipelines etc is
from the monitored field through wireless links; the data is common and can provide early warnings of damage to these
forwarded, possibly via multiple hops relaying, to a sink that
structures. Such early warnings help in minimizing loss of
can use it locally, or is connected to other networks (e.g., the
property and also human life.
Internet) through a gateway.
Rest of the paper is organized as follows: Section 2 pre- 2.1. Military Applications
sents the major applications of WSNs, the node and net-
work architecture is presented in section 3. Design chal- Wireless sensor networks can be an integral part of mili-
tary Command, Control, Communications, Computing, Intel-
lenges and communication architecture is presented in sec-
ligence, Surveillance, Reconnaissance and Targeting
tion 4 and 5 respectively. Routing challenges with static
(C4ISRT) systems [3]. During cold war era acoustic net-
and mobile sink are presented in section 6. Section 7 pre-
works were deployed in the United States for submarine sur-
sents the research challenges being faced. Finally, the con-
veillance [4]. Few of these sensors are still used to monitor
clusion is in section 8.
seismic activity in the region. Military C4ISRT prescribes
rapid deployment, self-organization and fault tolerance char-
2. APPLICATIONS OF WSNS
acteristics for sensing. WSNs have all these characteristics
Today, sensors capable of observing different physical making them a very promising sensing technique for military
phenomenon have been designed successfully. Proliferation C4ISRT. Sensor networks are densely deployed, have low
of Micro-Electro-Mechanical Systems (MEMS) technology costs of deployment and failure of few sensors does not af-
has made design of smart sensors possible. Sensors are now fect the performance. This makes the concept of deploying
capable of observing not only ambient conditions like tem- sensor networks highly suitable for battlefields. However,
perature, pressure flow etc. but can also be deployed for ob- such applications require protection against rouge sensors.
serving lot of others phenomenon. They can sense the envi- Some of the military applications of sensor networks are
ronment, measure and send data wirelessly to control unit for monitoring friendly forces, equipment and ammunition; bat-
tlefield surveillance; reconnaissance of opposing forces and
terrain; targeting; battle damage assessment; and Nuclear,
*Address correspondence to this author at the Department of Mathematics,
Government College, Kota, India; E-mail: arunkr71@gmail.com Biological and Chemical (NBC) attack detection and recon-
naissance [3].

2210-3287/13 $58.00+.00 © 2013 Bentham Science Publishers

26 International Journal of Sensors, Wireless Communications and Control, 2013, Vol. 3, No. 1 Gupta and Kumar

2.2. Environmental Applications width. These medical sensors (e.g., ECG electrodes, pulse
oximeter, blood pressure, and temperature sensors) are de-
Concerns for environment are increasing with every pass-
ployed on patient’s body and collect the individual’s physio-
ing day. Environmental Scientists are working hard to find
logical data and sends the collected data via a wireless chan-
the impact of industrial and other activities on the environ-
nel to health professionals’ hand-held devices (i.e., PDA,
ment particularly the climate change. Apart from monitoring iPhone, laptop, etc) [5].
environment, monitoring wild life, aqua life, movement of
birds are few other concerns for environmentalists. Remote Wireless medical sensor technology has offered tremen-
sensing satellites help in collecting data for prediction on dous advantages to healthcare applications, such as continu-
weather, rainfall, thunderstorms and also for detection of ous patient monitoring, mass-causality disaster monitoring,
forest fires. Monitoring of health of soil and moisture are large-scale in-field medical monitoring, emergency response,
important for good agricultural yields. Wireless Sensors have etc. Further, these WMSNs provide many new ways for
been deployed extensively for such applications [3]. acute disease analysis (e.g., motion analysis for Parkinson’s
disease).
Deployment of sensors for environment monitoring re-
quires integration of information across temporal, spatial and Pardeep Kumar et al. [5] studied the WMSNs and re-
spectral scales. Further, sensors for observing environment quirements of making these networks secure. Reliable data
are required to be deployed in remote areas in air, on surface transmission, fast event detection, timely delivery of data,
and also underwater. Such deployment prescribes for unat- power management, node computation and middleware, pa-
tended operation for long duration. Hence, sometimes it may tients’ security and privacy are the big concerns for
be necessary to equip the sensors with power scavenging healthcare applications. The physiological data of an indi-
systems like solar cells. The communication systems are vidual are highly vulnerable. With sensors in WMSNs being
required to be functional under different medium and also in wireless, the network can be eavesdropped easily even with
the presence of obstructions. As the attenuation of radio sig- common devices like smart phones, iPhones, PDAs and lap-
nals is dependent upon the environmental conditions; care tops [6]. This may cause unauthorized collection of patient’s
needs to taken to ensure communication under all weather data by eavesdropper resulting in loss of patient’s privacy. In
conditions. some countries, disclosure of medical data and information
on patients are highly regulated (e.g., the Health Insurance
2.3. Health Applications Portability and Accountability Act of 1996 (HIPAA), USA).
Unauthorized disclosure may result in penalties on
WSNs are widely deployed in health care industry. These healthcare providers. Furthermore, as wireless medical sen-
networks, more commonly termed as, Wireless Medical Sen- sor nodes themselves provide services to users (doctors,
sor Networks (WMSN), are now cutting-edge component of nurses, and technicians, are a few examples) it is necessary
health-care industry and are capable of improving quality-of- to control and regulate access to these sensors. Therefore,
care to the patient without sacrificing the comfort. WMSN strong user authentication is a core requirement to protect
consists of lightweight devices with limited memory, low from illegal access to patients’ vital signs, and can attain the
computation processing, low-battery power and low band-

Fig. (1). Typical wireless medical sensor network in a hospital environment.

Wireless Sensor Networks International Journal of Sensors, Wireless Communications and Control, 2013, Vol. 3, No. 1 27

highest levels of patients’ privacy. E-SAP [5] is a strong strain gauges, anemometers, weigh-in-motion devices and
authentication protocol and includes: temperature sensors.
1). A two-factor (i.e., password and smartcard) professional Wireless sensors can also be used to monitor the state of
authentication; road surfaces. For example, the Massachusetts Institute of
Technology (MIT) carried out a research project to detect the
2). Mutual authentication between the professional and the
number of potholes in a road, using Boston taxis to cover the
medical sensor;
entire city. A similar approach was undertaken by the Uni-
3). Symmetric encryption/decryption for providing message versity of Sri Lanka to monitor Sri Lanka´s roads.
confidentiality;
Additionally, monitoring systems in tunnels are also
4). Establishment of a secure session key at the end of widespread around the world. From air flow to visibility, and
authentication; a wide range of gases (CO, CO2, NO2, O2, H2S and PM-10)
5). Change of password used by professional i.e. the doctors, are the most demanded parameters to monitor air quality
nurse etc. inside tunnels. At this time, many of these systems are wired
installations: the deployment of Wireless Sensor Networks
For healthcare, deployment of wearable sensors is gain- would save money, increase safety and reduce installation
ing momentum [7]. The wearable sensors need to be ultra times.
lightweight and shall consume very little power. Further, the
patient may move out of the radio range. Hence, it should be Apart from the above applications, researchers have de-
possible to store the data on the sensor. The network must veloped and deployed WSNs for specific applications like
tune its operation in the face of variations in radio band- Improve Road Monitoring [1], Electricity Substation Moni-
width (e.g., as the patient moves around or leaves the home) toring [1], pipeline infrastructure monitoring [9], habitat
and energy availability (based on the activity level of each monitoring [10], railway infrastructure monitoring [11, 12].
sensor). Most importantly, the system must yield high- Flouri [13] presents versatile software architecture for civil
quality, clinically-relevant data. The high data rates of the structure monitoring with wireless sensor networks. More
sensors involved (100s of Hz per channel or more) demand and more applications employing WSNs are being built [14-16].
on-board processing and storage. It is infeasible to transmit
the complete sensor data because it would rapidly deplete the 3. ARCHITECTURE OF WSN
nodes’ batteries. The batteries should last for days and shall Fig. (2) presents architecture of a typical sensor node. A
not require frequent charging. typical sensor unit is required to be of small size (a coat but-
ton) and has on-board sensors, wireless transceivers, a small
2.4. Infrastructure Monitoring processor and a battery. The nodes may be stationery or
Infrastructure development is the key to the growth of moving, may or may not be aware of their location, may be
every society. Cost of building infrastructures like, roads, homogeneous or heterogeneous and may or may not have a
bridges, dams, pipelines, airports etc. are huge and spread power generation unit. However, in our study, we will be
over long durations. Monitoring the utilization and health of considering that nodes are (i) stationery, (ii) unaware of their
such systems is vital for economic growth. With the threats location (iii) homogeneous and (iv) do not have capability to
due to terror activities ever rising, it is critical to protect such recharge the power unit comprising mainly of battery. The
key components of national infrastructure. Further, even sensor produces analogue signals based on the observed
during construction of such projects, accidents are caused phenomenon. These signals are converted to digital data by
due to design faults/ defects human errors. Such accidents the ADC and fed to the processing unit. The processing unit,
cause loss of human life and it is important to get early warn- which generally has an on board storage in small quantity,
ing signals to avoid accidents. Under construction bridge on manages the procedures that make the sensor node collabo-
the Chambal River in Kota (Rajasthan) collapsed in Dec. rate with the other nodes to carry out the assigned sensing
2009 claiming 48 lives. It is possible to at least get early sig- tasks. A transceiver unit connects the node to other nodes
nals of impending failure of such accidents using WSN tech- and also with the Base Station thus forming the WSN. The
nology. nodes, when sprayed in a targeted area, are capable of form-
ing a communication network on the fly.
Throughout the years, many transport infrastructures-
bridges, tunnels or viaducts-have collapsed due to natural Sensor nodes are generally deployed in irregular topol-
disasters or because of poor maintenance. One of the best ogy and are densely populated as shown in the (Fig. 3). The
examples is the bridge in Minneapolis in 2007 that killed 13 remote sensors (also termed as Motes) sense the phenome-
people and injured 145. In 2008, this bridge was re-built us- non, pre-process it after converting the sensed quantity into
ing a sensing system to collect data regarding structural be- data and transmit it further. The observed phenomenon is
havior and corrosion [8]. Monitoring bridges is one of the finally processed at a central location having sufficient proc-
more successful applications of Smart Roads. For instance, essing capabilities known as Base Station (BS). Data tra-
the six-lane, 2.9 km Charilaos Trikoupis Bridge in Greece is versal from motes to the base station requires a path to be
outfitted with 100 sensors that monitor its condition. Soon established. This path is generally multi-hop. Depending
after opening in 2004, the sensors detected abnormal vibra- upon the routing algorithm, the path may use one of the sen-
tions in the cables holding the bridge, which led engineers to sor nodes as head of the cluster of sensors and may transmit
install additional weight to dampen the cables. The sensor the data along the path. However, finding the path from sen-
networks for these kinds of bridges include accelerometers,
28 International Journal of Sensors, Wireless Communications and Control, 2013, Vol. 3, No. 1 Gupta and Kumar

sor to the base station is a challenge due to reasons presented power consumption. Synchronization and localization are
in subsequent sections. required (i) to support TDMA, (ii) to determine temporal
ordering of messages (iii) to determine proximity of sensors,
(iv) to determine relative position of sensors with respect to
each other and (v) to determine the exact time of observed
event [4-18]. Importance of Data Aggregation and Data Dis-
semination has been highlighted by Rajgopal et al. [19]. Se-
curity in WSNs has been investigated by several researchers
[20-22]. Design of operating system and middleware for
WSNs requires special attention due to limited storage and
limited computing resources at the node. Yang et al. [23]
studied the issues in the design of middleware in detail.
Apart from these general issues, every application of
WSNs presents new challenges and studies have been carried
out to address these application specific issues. Ameen et al.
[24] studied security issues in employing WSNs for
healthcare. Gungor et al. [25, 26] presents the challenges in
smart grid and industrial applications. Underwater deploy-
ment of WSNs require using acoustic signals for communi-
Fig. (2). Sensor Node architecture. cation rather than radio signals and undersea currents affect
the placement of sensors. These challenges in underwater
deployment have been studied by Heidemann, John, et al.
4. WIRELESS SENSOR NETWORKS: DESIGN
[15]. As the pipelines traverse long stretches of land, design
CHALLENGES
of networks for pipeline monitoring requires that the sensors
Fig. (2) presents a schematic of a sensor node. The gen- gather data along the length and send them to respective con-
eral arrangement of sensors in WSN is shown in (Fig. 3). trol point. The collected data and measurements are then sent
Realization of a successful WSN thus requires proper plan- back to a central control station.
ning and careful design. Deployment of WSNs requires In all the WSNs irrespective of their application, energy
knowledge of communication and signal processing, hard- is the most important constraint. The nodes in WSNs operate
ware technologies, embedded system design and software on batteries. In most of the applications, it is not possible to
engineering. These areas of technologies are vastly different recharge the batteries after deployment. Only in few cases,
from each other and thus a in depth study of influence of one energy scavenging systems are possible to provide (dashed
factor on the design of other factors is required. Here we lines in (Fig. 2) indicates this fact). Two factors together
present a brief survey of the factors affecting design and per- mandate that the sensor consume as little power as possible.
formance of WSNs.
Now we present each of these issues and their effects
together with efforts to overcome these challenges.

4.1. Scalability
The number of nodes deployed in an application is de-
termined by the requirements of application. The density
may range from few nodes to thousands of nodes depending
upon the application. Communication and data aggregation
algorithms are required to works under varying densities. In
low density networks, only few nodes may send the data
whereas sensing of data by multiple nodes in densely popu-
lated networks may produce multiple data values of the same
phenomenon at a given time. The data aggregation algo-
rithms need to be capable of identifying duplicity of data.
Further, multiple sensors sending the same data values drain
the batteries as well.

4.2. Communication
Sensor networks have unique set of challenges for net-
working. Deployment of WSNs for an application can be
justified only if it continues collecting and sending data for
Fig. (3). Typical Sensor Network. longer stipulated life. Research has shown that radio com-
munication is a major consumer of energy in WSNs. A com-
As presented in [3-17], major design constraints are parison of the cost of computation to communication in fu-
hardware constraints, fault tolerance, scalability, production ture platforms by Pottie and Kaiser presented by Ganeshan
costs, sensor, network topology, transmission media; and
Wireless Sensor Networks International Journal of Sensors, Wireless Communications and Control, 2013, Vol. 3, No. 1 29

[17] reveals that 3000 instructions can be executed for the • Consume extremely low power,
same cost as the transmission of one bit over 100 m. The
• Operate in high volumetric densities,
power loss in general is inversely proportional to the square
of distance. However, in case of WSNs, the power loss is • Have low production cost and be dispensable,
generally of fourth order [3]. This observation suggests that • Be autonomous and operate unattended,
the communication protocols be designed so as to consume
as little energy as possible. Energy conservation in such net- • Be adaptive to the environment.
works employs two approaches to minimize communication The communication circuitry though requires low duty
over-head. The first is at the MAC and networking layers, cycle; designing energy efficient low duty cycle communica-
where nodes turn off their radios when they are not required tion circuit is a challenge. Further, in spite of availability of
for communication (adaptive duty cycling) [4-17]. The sec- more computational power in small size processors, process-
ond is data reduction through in-network processing (also ing power and memory are still scare resource in a node.
called data aggregation), whereby correlations in data are
exploited [17] to reduce the size of data, and correspondingly 4.5. Sensor Network Topology
communication cost. As presented by Akyildiz [3], the total
radio power consumption Pc is: Topology maintenance in WSNs is challenging because
of large number of sensor nodes and also frequent failures
Pc= N T(PT(Ton+Tst) + Pout(Ton)) + NR[PR(Ron+Rst)] either due to power depletion or destruction. It is generally
where, PT/R is power consumed by the transmitter/receiver; not possible to engineer the topology before deployment and
Pout, the output power of the transmitter; Ton/Ron, the trans- thus the data gathering and communication requires working
mitter/receiver on time; Tst/Rst, the transmitter/receiver start- under irregular and dynamically changing topology.
up time and NT/R, the number of times transmitter/receiver is
switched on per unit time, which depends on the task and 4.6. Environment
medium access control (MAC) scheme used.
WSNs may be deployed in varying environmental and
Further, hidden node and exposed node problems also ambient conditions. These may be required to operate under-
need to be addressed at MAC layer [4]. Ganeshan [17] pre- sea, in desert, in extreme cold conditions, on road intersec-
sents four techniques to address the communication chal- tions, may be tied to animals (Project Tiger) and so on.
lenges. Every environment is different and the nodes need to sense
the phenomenon and the communication systems shall work.
4.3. Data Collection and Dissemination This requires a careful study of the environment on the per-
formance of the network.
In WSNs, each sensor collects data periodically and
transmits it to the BS. Alternatively, the collected data may
4.7. Transmission Media
be sent to a cluster head. The frequency of data collection
and its transmission is application dependent. Data gathering The wide range of applications of sensor networks makes
is defined as the systematic collection of sensed data from the choice of transmission media more challenging. Radio
multiple sensors to be eventually transmitted to the base sta- Communication is the main transmission media in WSNs.
tion for processing [19]. Transmission of data by all the sen- However, selection of frequency for communication faces
sors to the BS is inefficient and puts additional burden on challenges due to regulations on frequency spectrum. These
already energy deficient sensor nodes. Data from neighbour- regulations suggest use of ISM bands for communication
ing nodes being highly correlated and redundant, it is not which are unregulated. Further, the size of antenna and
necessary to transmit data from all the nodes. Data aggrega- power consumption by communication circuitry within the
tion techniques are employed to process the data either at the node is highly dependent on the frequency. It suggests that
sensor or at an intermediate node to reduce the number of the UHF band is better for these networks. 2.4 GHz band has
transmitted data packets to save energy. However, the data been found to suitable for these networks. This band pro-
aggregation algorithms are network topology dependent vides good energy efficiency and also small antenna size.
which affects routing and also closely related the data gather- The unregulated nature of ISM band causes difficulties be-
ing at the nodes [27]. Rajgopalan [19] also presents a survey cause the same frequency can be used by other applications
of data aggregation algorithms for different network topolo- as well resulting in radio interference.
gies and architectures.
Infrared which is license free and not affected by inter-
ference is the alternate to radio communication. Infrared-
4.4. Hardware Constraints based transceivers are cheaper and easier to build. However,
A node consists of four major units; sensing unit, a proc- the requirement of Line of Sight between the sender and re-
essing unit, a transceiver unit and a power unit. These four ceiver makes it a reluctant choice for sensor networks. Nev-
units are required to be packed in a match box size casing. ertheless, infrared can be used in harsh environments, where
However, most of the applications require the size to be even RF signals suffer from high attenuation, such as underwater
smaller. Apart from limitation on the size, the nodes are re- links [3].
quired to be lightweight. In some cases, the size may be less Akyildiz [3] reports use of optical communication in
than a cubic centimeter making it possible to remain sus- WSNs. Two transmission schemes, passive transmission
pended in the air. However, size and weight are not the only using a corner-cube retro reflector (CCR), and active com-
constraints on sensor nodes. The nodes are required to [3]: munication using a laser diode and steerable mirrors, are
30 International Journal of Sensors, Wireless Communications and Control, 2013, Vol. 3, No. 1 Gupta and Kumar

reported to have been examined by researchers. In the for- sible for receiving the data sent by its neighbors and for-
mer, the mote does not require an onboard light source. A warding these data to one of its neighbors according to the
configuration of three mirrors (CCR) is used to communicate routing decision. Considering the importance of power con-
a digital high or low. The latter uses an onboard laser diode sumption, researchers have suggested making the design
and an active-steered laser communication system to send a energy aware. The survey is presented in [29].
tightly collimated light beam toward the intended receiver.
Anastasi et al. [29] presented survey on energy conserva-
Marine applications require using long wavelength radia- tion in WSNs. It has been found that energy consumption in
tions capable of penetrating water surface as transmission transmission is much more than the energy consumed in
medium [3-15]. Inhospitable terrain or battlefield applica- sensing and processing [1, 3, 29, 30]. Considering this fact,
tions might encounter error-prone channels and greater inter- research has been focused on reducing the power consump-
ference. Moreover, a sensor antenna might not have the tion in transmission of data. This will be presented sepa-
height and radiation power of those in other wireless devices. rately. The lifetime of a sensor network can be extended by
Hence, the choice of transmission medium must be sup- jointly applying different techniques for minimizing power
ported by robust coding and modulation schemes that effi- consumption in transmission as well as computation. For
ciently model these vastly different channel characteristics. example, energy efficient protocols are aimed at minimizing
Considering these facts, acoustic communication techniques the energy consumption during network activities. However,
have recently been adopted for underwater sensor network a large amount of energy is consumed by node components
applications instead of RF waves [15], which experience (CPU, radio, etc.) even if they are idle. Power management
high attenuation in this environment. schemes are thus used for switching off node components
that are not temporarily needed [30].
Ultra Wide Band (UWB) which is a particular carrier-
free spread spectrum technique where the RF signal is spread Holland, Matthew, et al. [31] presents a metric called the
over a spectrum as large as several GHz; is of great interest energy per successfully received bit, which specifies the ex-
for sensor networks since it meets some of their main re- pected energy required to transmit a bit successfully over a
quirements. This implies that UWB signals look like noise to particular distance given a channel noise model. By minimiz-
conventional radios. Such signals are produced using base- ing this metric, for different modulation schemes, the en-
band pulses (for instance, Gaussian mono-pulses) whose ergy-optimal relay distance and the optimal transmit energy
length ranges from 100 ps to 1 ns, and baseband transmis- as a function of channel noise level and path loss exponent
sion is generally carried out by means of pulse position can be found. This enables network designers to select the
modulation (PPM). Modulation and demodulation are indeed hop distance, transmit power and/or modulation scheme that
extremely cheap. UWB provides built-in ranging capabilities maximize network lifetime.
(a wideband signal allows a good time resolution and there-
fore good location accuracy), allows very low power con- 5. NETWORKING ARCHITECTURE FOR WSNS
sumption, and performs well in the presence of multipath
fading [28]. The sensor nodes are usually scattered in a sensor field.
Each of these scattered sensor nodes has the capabilities to
collect data and route data back to the sink and the end users.
4.8. Power Consumption
Data are routed back to the end user by a multi-hop infra-
In WSNs, power consumption is the key concern. As structure less architecture through the sink as shown in (Fig. 4).
shown in (Fig. 2), nodes are battery powered. Due to cost The sink may communicate with the task manager node via
and size constraints, the battery provides a limited power. In Internet.
most of the cases, it is not possible to replenish or replace the
battery once it has drained out. In the event of draining out of
battery, the node stops working and considered to be dead.
Increased population of dead nodes reduces the data collec-
tion and transmission of data to the BS. This leaves the net-
work unsuitable for the application. It suggests that while
designing the WSNs or any component thereof, care needs to
be taken to consume as little energy as possible.
In multi-hop ad hoc sensor networks, each node plays
two separate and complementary roles; Data Originator and
Data Router. As data originator, each node is required to
gather data from the on board sensors, process it and transmit
it to the neighboring node for transmission to the BS.
In addition to originating data, each sensor node is re-
sponsible for relaying the information transmitted by its Fig. (4). WSN Deployment and Communication Architecture.
neighbors. The low-power communication techniques in
WSNs limit the communication range of a node. In a large Fig. (5) shows the protocol stack for WSNs. This proto-
network, multi-hop communication is required so that nodes col stack combines power and routing awareness, integrates
relay the information sent by their neighbors to the data col- data with networking protocols, communicates power effi-
lector, i.e., the sink. Accordingly, the sensor node is respon- ciently through the wireless medium, and promotes coopera-
tive efforts of sensor nodes. The tasks performed by each
Wireless Sensor Networks International Journal of Sensors, Wireless Communications and Control, 2013, Vol. 3, No. 1 31

layer are presented in (Table 1). Power, mobility, and task is not possible to recharge the node batteries; the algorithms
management planes monitor the power, movement, and task designed for Mobile Ad hoc Networks (MANET) are unsuit-
distribution among the sensor nodes. These planes help the able for WSNs.
sensor nodes coordinate the sensing task and lower the over-
WSNs pose unique characteristics such as denser level of
all power consumption. node deployment, higher unreliability of sensor nodes and
severe energy, computation, and storage constraints. These
characteristics make routing in WSNs a challenging task.
The routing algorithm is also required to overcome the fol-
lowing challenges:
1). It is not possible to build a global addressing scheme for
the deployment of sheer number of sensor nodes. There-
fore, classical IP-based protocols cannot be applied to
sensor networks.
2). Furthermore, sensor nodes that are deployed in an ad
hoc manner need to be self-organizing as the ad hoc de-
ployment of these nodes requires the system to form
connections and cope with the resultant nodal distribu-
tion, especially as the operation of sensor networks is
unattended.
3). In contrary to typical communication networks almost
all applications of sensor networks require the flow of
Fig. (5). The sensor networks protocol stack. sensed data from multiple regions (sources) to a particu-
lar sink (Base Station). This, however, does not prevent
the flow of data to be in other forms (e.g., multicast or
Table 1. Tasks Performed by Layers.
peer to peer).
4). In most application scenarios, nodes in WSNs are gen-
Layer/Plane Major Tasks Performed erally stationary after deployment except for maybe a
few mobile nodes. Nodes in other traditional wireless
Application Application hosting and user interaction networks are free to move, which results in unpredict-
able and frequent topological changes. However, in
Transport Maintaining data flow
some applications, some sensor nodes may be allowed
Network Routing of Data to move and change their location (although with very
low mobility).
Power Management, minimize collision with
Data Link 5). Generated data traffic has significant redundancy in it
neighbors’ broadcast
since multiple sensors may generate same data within
Physical Modulation, transmission and reception the vicinity of a phenomenon. Such redundancy needs to
be exploited by the routing protocols to improve energy
Design of protocol for each of these layers poses chal- and bandwidth utilization.
lenges specific to the layer. Our work being related to rout- 6). Sensor nodes are tightly constrained in terms of trans-
ing we will be focusing on the challenges in the design of mission power, on-board energy, processing capacity
routing protocols. Issues and challenges associated with and storage and thus require careful resource manage-
other layers are presented in [3, 15, 28, 31, 32]. ment.
6. DESIGN CHALLENGES IN ROUTING PROTO- 7). Sensor networks are application-specific (i.e., design
COLS requirements of a sensor network change with applica-
tion). For example, the challenging problem of low-
6.1. With Stationary Base Station latency precision tactical surveillance is different from
At the network layer, the main aim is to design for en- that of a periodic weather monitoring task.
ergy-efficient route setup and reliable relaying of data from 8). Nodes in the network are always randomly deployed and
the sensor nodes to the sink so that the lifetime of the net- the position information is not available without a
work is maximized. Choosing the same route for every Global Positioning System (GPS) service for the sake of
communication depletes the energy of intermediate nodes economic cost reduction.
quickly causing their death. Such death of nodes will make
the network unusable. Consequently, it is desired that the Cluster based approach has been suggested by several
routing algorithm considers the available energy with the researchers. However, such approach requires re-election of
node while deciding the routes. Furthermore, due to chang- cluster head to preserve the energy at the cluster head. Simi-
ing topology and large number of nodes with limited data larly, fixed location of base station or a single base station
storage capabilities, it is not recommended to store the rout- causes the nodes close to BS die out quicker than rest of the
ing tables at the nodes. As presented earlier, in most cases it nodes. The phenomenon known as “energy hole problem
32 International Journal of Sensors, Wireless Communications and Control, 2013, Vol. 3, No. 1 Gupta and Kumar

[33]” makes the network unusable in spite of majority of tern to further optimize the detection of mobile sinks. In
nodes being still live. These characteristics of WSNs suggest fact, if visiting times are known or can be predicted with
that routing algorithms be adaptive, robust and self healing certain accuracy (e.g. sinks mounted on public transport
and should work in a localized way [34-37]. buses/trams plying through city streets collecting pollu-
tion data), sensor nodes can be awake only when they
6.2. With Mobile Base Station expect the mobile sink to be in their transmission range
[46].
Data collection from sensors through stationary Base
Stations located at fixed locations increases the duty cycle of 4). Mobility Planning: For optimal performance, the path or
nodes closer to the BS. This causes the energy of these nodes trajectory of the sink as well as the speed need to be
to deplete at a much faster rate as compared to nodes away carefully planned. The sink may move around at con-
from the BS. This reduces the lifetime of the network due to stant speed or may make moves to fixed locations.
early death of sensors close to the BS [38-45]. This has However, finding an optimal tour or trajectory is a chal-
prompted the researchers to investigate using mobile sinks lenging task. The sink may move in random or prede-
for data collection. In few applications like wildlife monitor- termined trajectory or may compute its trajectory adap-
ing, even a three tier structure with the mobile sink passively tively [46, 47].
listening to incoming data from sensors has been suggested. a). Random mobility can result in incomplete data
The data collected by the mobile sink is then transmitted to collection from the WSN, because with a random
the fixed BS [38]. mobility pattern there is no guarantee that the sink
Luo et al. [39] investigated the routing towards mobile will reach all nodes in the sensor field or it might
sinks. They have found that using mobile sinks have a posi- take too much time to do so. If the time required
tive effect on the lifetime of the networks. for complete coverage of the field has to be even
lower, then the sink can be programmed to collect
Mobility in WSNs can be employed in two different
data from all nodes which are within a maximum
ways:
number of hops larger than one. This results in in-
I. Few of the nodes are equipped to be mobile. Such nodes creased relaying load on the sensor nodes.
have either more powerful batteries or have the power
b). With fixed mobility, the path is predetermined and
scavenging capability. These nodes move in the de- is not influenced by the behaviour of the WSN at
ployment and collects data from the nodes. The sensed
runtime. Coverage of the sensor field has to be
data is then forwarded to the sink.
guaranteed by an appropriate strategy for deter-
II. One or more mobile sinks are deployed to collect the mining the routing paths for the data packets. An
sensed data. These sinks move around the field and col- important distinction is whether the sink can pre-
lect data. The sink mobility may either be controlled or dict its future positions or not. If it can’t, reactive
uncontrolled. In uncontrolled mobility, the sink is data forwarding mechanism using a pull strategy
mounted on certain mobile entity such as shuttle bus; the based on request messages broadcasted by the sink
movement of which can not be controlled by network. In needs to be employed. Moreover, sink mobility
controlled mobility, the sink is an integral part of the needs to be planned such that the complete sensor
network and is thus fully controlled. field is traversed in minimum possible time mini-
Using mobile sinks or deploying nodes with mobility mizing energy dissipation. In case the sink is able
to predict its future positions it can communicate
pose similar challenges different from the challenges present
this information to a node located in the vicinity of
with static sinks. These challenges are:
its future position. This node is responsible for
1). Cost: Energy consumption and cost of mobile nodes/ collecting the sensor data in its vicinity so that
sink is higher than static nodes/ static sinks. Cost also when the sink actually arrives at this position, it
increases at a power function of the speed at which the should not have to wait for the data.
mobile sink moves around. Hence, deploying large
c). Controlled mobility refers to schemes where sink
number of mobile nodes may not be practical. Similarly,
mobility is controlled or guided based on a pa-
deploying sinks moving at high speed may not be eco-
rameter of interest, such as residual energy of the
nomical.
nodes, or on a predefined objective function, or on
2). Data Collection Strategy: With static sinks, generally, predefined observable events.
the sensed data is transmitted (Pushed) by the nodes to-
5). Security: Due to the operating nature of WSNs, these are
wards the sink either in a single hop or multi-hop com-
often left unattended, hence prone to different kinds of
munication. With mobile sinks, the sink needs to reach
malicious attacks such as the Sybil attacks, clone at-
the node to collect (Pull) the sensed data. This requires
tacks, node replication attacks, mobile sink replication
nodes to be aware of the time at which the sink is at the
shortest distance and schedule the data transmission ac- attacks, and wormhole attacks. Thus, security services,
(such as authentication and pair wise key establishment),
cordingly. Because of this, deployment of mobile sinks
and countermeasure attacks are vital [49].
in event driven networks requires special considerations
[46, 47, 48]. 6). Sojourn Locations and sojourn times [50]: One approach
3). Mobility-aware power management: In some cases, it is for achieving improved network lifetime is to make the
possible to exploit the knowledge on the mobility pat- sink visit predetermined locations within the observed
Wireless Sensor Networks International Journal of Sensors, Wireless Communications and Control, 2013, Vol. 3, No. 1 33

area. These locations are called rendezvous points or so- However, future applications may require solution to the
journ locations. Finding optimal sojourn locations is a following issues:
non-trivial task and requires considering the residual en-
ergy in the node(s) close to sojourn location. However, 7.1. Quality of Service
optimizing sojourn times defined as the duration for
Throughput, delay, correctness of measured phenome-
which the sink stays at a sojourn is even more challeng- non, transmission errors even with unreliable communication
ing. Shorter duration may result in incomplete data col- links and node failure and lost data packets are important
lection whereas longer sojourn time may make collected QoS parameters. However, in WSNs, achieving QoS is diffi-
data irrelevant due to excessive delays. cult due to constrained energy and computation capabilities.
7). Location Update [51-53]: In fixed networks, the sensing Further, the required QoS is highly application dependent. In
nodes are aware of the location of the sink. However, event driven systems, delivering the data reliably and with-
with mobile sinks the nodes may need to find the loca- out errors within acceptable delay is more important than
tion of sink before transmitting the sensed phenomenon. sending periodic information of no event. Improving
Locating the sink by nodes is difficult due to constrained throughput is another important consideration. Single path
resources. Frequent location updates from mobile sinks routing limits the achievable network throughput. For appli-
can lead to both rapid energy consumption and increased cation requiring high throughput, routing protocols employ-
collision in wireless transmission. Various methods such ing multiple paths are better suited. Real time applications
as deploying special nodes called locators, using the require time bound data delivery. A survey of QoS aware
overhearing capabilities of nodes in wireless channels, routing is presented in [54].
broadcast in limited regions are few of the approaches
for location update. 7.2. Security
8). Number of sinks: When the observed field is too large to Providing security in WSNs is a challenge due to use of
accept the delay with single sink, multiple sinks may be wireless communication, low computing and energy re-
deployed. However, with multiple mobile sinks, the tra- sources, tiny size of nodes, scattered deployment of nodes
jectories of all the sinks need to be controlled in such a which makes the topology unknown, and high risk of physi-
way that the transmitting node is visited only by one cal attacks to unattended sensors. WSNs require confidenti-
sink in every round. ality, integrity, authenticity, and availability to make their
operations secure [20-22]. Unlike traditional networks, sen-
9). Delay Tolerance: In event driven networks, the sensed sor nodes are deployed physically in open areas where there
data may be required to be transmitted without delay. is added risk of intervention with people and environment.
However, with mobile sinks, it is possible that the sink However, limited computing resources and low energy
may not be at optimal location at the time of event or budget make the conventional schemes unsuitable for these
even may not be available at all. This introduces delays networks. Further, the security requirements vary from ap-
in transmission. Nodes may require storing the data. plication to application.
Thus, the data collection strategy is application depend-
Almost all security mechanisms use keys for providing
ent.
security. Management of keys involves generation and dis-
10). Reliability of networks measured in terms of packet tribution of keys and also revocation of compromised keys.
delivery ratio may be low. Further, use of keys increase computations resulting in in-
Fig. (6) presents the methods of data gathering in WSNs creased energy consumption. All these factors suggest that
with mobile sinks. In Delay Tolerant WSNs, direct contact the security threats and protection against these threats in
WSNs need to be addressed differently.
data collection requiring the mobile sink to reach the point of
sensing results in maximum energy savings. The sink may Mostly, sensor nodes are deployed in unattended fields
either visit each data source to obtain data directly or may and wireless channel is used for communication. The net-
visit few selected rendezvous points (RPs). In the former works are thus prone to rogue nodes placement and channels
case, the overhead due to message relay is completely elimi- can be jammed. A spurious node may continue to transmit
nated. However, it has large data collection latency for the causing excessive collisions. Further, the data may be inter-
slow moving sinks. In the later, data is buffered and is ag- cepted and/or modified on the wireless channel. Security
gregated at RPs avoiding long travel distance by using multi- solutions based on long cryptographic keys and requiring
hop data communication. frequent replacement/ exchange of keys are not suitable due
to energy and computing constraints. Use of the same key
In WSNs requiring data collection in real time, the sink is across the network is risky as the only one compromised
required to be adaptively relocated to positions with largest node can reveal the key. Current solutions like establishing
energy gains. Thus, optimal location for mobile sink is re- pair wise keys with neighbours, selecting keys from a pool of
quired to be computed and data needs to be routed to the pre-defined keys are also impractical due to reasons such as
optimal location adaptively. node failure, large number of node pairs etc. Spread spec-
trum techniques generally employed for security against
7. RESEARCH CHALLENGES IN WSNS jamming attacks are (i) energy intensive and (ii) require
Considerable progress has been made in designing cost wider spectrum. With the number of sensors being very
effective solutions for vast range of applications of WSNs. large, frequency hoping is impractical. Though, the comput-
ing constraints may be overcome, increasing computing ca-
34 International Journal of Sensors, Wireless Communications and Control, 2013, Vol. 3, No. 1 Gupta and Kumar

Fig. (6). Taxonomy of energy-efficient data gathering by mobile sinks [49].

pacity requires more energy. A detailed survey of possible ings in the observed data. Methods for detection of such out-
threats in WSNs at each layer is presented in [55]. However, lier data needs to be developed. However, such outlier data
security is a continuous process as new attacks will be de- may also be generated due to intrusion into the network. It is
signed by malicious users. therefore necessary to distinguish whether the outlier data
has been generated due to battery mal function or intrusion
7.3. Fault Tolerance [58, 59]. In case, data aggregation is performed within the
sensor network, the incorrect data from few sensors may
Due to factors such as physical damage, draining of bat- cause a data message containing incorrect data. This needs to
tery, environmental interference etc. few of the sensing be addressed within the network and can not be left to the sink.
nodes may stop functioning. It requires that the WSNs
should have capabilities of continuing to function even if few Fault detection and fault recovery techniques are other
of the nodes fail. The communication protocol, in particular, major challenges. Self diagnosis, group detection, and hier-
should be capable of sending data with reduced number of archical detection are few of the techniques for fault detec-
nodes. However, the required fault tolerance is highly de- tion. Fault recovery techniques include passive replication
pendent upon the requirements of application and occurrence and service distribution once a failure has been detected.
Providing redundancy for fault recovery may not be eco-
of fault is affected by the environment in which the network
nomically viable.
is deployed. The reliability Rk(t) for fault tolerance of a sen-
sor node is modelled using the Poisson distribution to cap- 7.4. Large Streaming Data Processing
ture the probability of not having a failure within the time
interval (0,t); Rk (t ) = e
k t In applications such as multimedia and ecological moni-
, where, k and t are the failure
toring, the generated data may be too large. To translate this
rate of sensor node k and the time period, respectively. ‘data deluge’ into scientific knowledge requires comparable
WSNs are often deployed and will operate in potentially advances in our ability to integrate, process and analyze
hostile, harsh and unconditioned environments. Further, the massive data sets [60]. The collected data is voluminous and
nodes and link are failure prone. The nodes may mal func- at times may be heterogeneous (may contain numerical val-
tion and send wrong data due to battery dropping below cer- ues, images etc.), and may contain errors caused by radio
tain levels, hardware and software bugs. Further, higher interference, faulty nodes. Detecting such errors in large data
structural complexities of nodes often results in higher fault sets is a challenging task. Further, storage of huge data set is
rates. also not easy. Relational Databases commonly used for stor-
ing data may not be suitable as redundancy is the essential
It is suggested that all the five primary types of re- characteristics of sensor data. Presenting these data in recog-
sources: computing, storage, communication, sensing and nizable and easy to understand formats is also necessary to
actuating can develop faults. However, each of these can interpret the sensed data. Suitable data mining techniques to
replace each other with suitable change in system and appli- get knowledge from the collected data need to be developed.
cation software. Sensing is most prone to faults [56]. How-
ever, the issue can be tackled by providing some redundant 8. CONCLUSION
sensors and changing the classification algorithm to adopt to
available resources. In this article the application areas of WSNs are dis-
cussed. The challenges in the design of sensors nodes and
Fault tolerant topology control is another solution sug- building the wireless communication infrastructure have also
gested by researchers. The emphasis is on maintaining k- been presented. Routing in WSNs is requires overcoming
vertex connectivity where any two nodes in the network can challenges of network life time due to energy constraints. It
maintain multiple disjoint paths [57]. However, the algo- has been suggested that use of mobile base stations will pro-
rithms need to be localized and shall ensure energy conserva- long the life of the network through uniform consumption of
tion while maintaining k-connectivity. stored energy which is a major constraint with stationary
When battery power falls below a certain level, the sen- base stations. Research challenges in designing WSNs for
sor readings may be incorrect. This generates outlier read- modern applications are also presented.
Wireless Sensor Networks International Journal of Sensors, Wireless Communications and Control, 2013, Vol. 3, No. 1 35

CONFLICT OF INTEREST [22] Bojkovic, Zoran S., Bojan M. Bakmaz, and Miodrag R. Bakmaz.
"Security issues in wireless sensor networks." International Journal
The authors confirm that this article content has no con- of Communications 2.1 (2008): 106-115.
flicts of interest. [23] Yu, Yang, Bhaskar Krishnamachari, and Viktor K. Prasanna. "Is-
sues in designing middleware for wireless sensor networks." Net-
work, IEEE 18.1 (2004): 15-21.
ACKNOWLEDGEMENTS [24] Al Ameen, Moshaddique, Jingwei Liu, and Kyungsup Kwak.
"Security and privacy issues in wireless sensor networks for
Declared none. healthcare applications." Journal of medical systems 36.1 (2012):
93-101.
REFERENCES [25] Gungor, Vehbi C., Bin Lu, and Gerhard P. Hancke. "Opportunities
and challenges of wireless sensor networks in smart grid." Indus-
[1] Mohammad Matin; Wireless Sensor Networks - Technology and trial Electronics, IEEE Transactions on 57.10 (2010): 3557-3564.
Applications, ISBN 978-953-51-0676-0, InTech, Published: July [26] Gungor, Vehbi C., and Gerhard P. Hancke. "Industrial wireless
18, 2012 (Edited Book). sensor networks: Challenges, design principles, and technical ap-
[2] Murthy, Manoj; Ad Hoc Wireless Networks: Architectures & proaches." Industrial Electronics, IEEE Transactions on 56.10
Protocols, ISBN 978-81-317-0688-6, Pearson Education, 2007. (2009): 4258-4265.
[3] Akyildiz, Ian F., et al. "Wireless sensor networks: a survey." Com- [27] Fasolo, Elena, et al. "In-network aggregation techniques for wire-
puter networks 38.4 (2002): 393-422. less sensor networks: a survey." Wireless Communications, IEEE
[4] Kazem Sohraby, Daniel Minoli, Taieb Znati; Wireless Sensor 14.2 (2007): 70-87.
Networks: Technology, Protocols, and Applications, ISBN 978-0- [28] Zhao, Yi Zhi, et al. "A survey and projection on medium access
471-74300-2, John Wiley & Sons, 2007. control protocols for wireless sensor networks." ACM Computing
[5] Pardeep Kumar, Sang-Gon Lee and Hoon-Jae Lee, E-SAP: Effi- Surveys (CSUR) 45.1 (2012): 7.
cient-Strong Authentication Protocol for Healthcare Applications [29] Anastasi, Giuseppe, et al. "Energy conservation in wireless sensor
Using Wireless Medical Sensor Networks, Sensors 2012, 12, pp networks: A survey." Ad Hoc Networks 7.3 (2009): 537-568.
1625-1647. [30] Daniele Puccinelli and Martin Haenggi, Wireless Sensor Net-
[6] Al Ameen, Moshaddique, Jingwei Liu, and Kyungsup Kwak. works: Applications and Challenges of Ubiquitous Sensing, IEEE
"Security and privacy issues in wireless sensor networks for Circuits and Systems Magazine, Third Quarter 2005, pp 19-29.
healthcare applications." Journal of medical systems 36.1 (2012): [31] Holland, Matthew, et al. "Optimizing physical-layer parameters for
93-101. wireless sensor networks." ACM Transactions on Sensor Networks
[7] Konrad Lorincz, et al. Mercury: A Wearable Sensor Network (TOSN) 7.4 (2011): 28.
Platform for High-Fidelity Motion Analysis, Proceedings of the [32] Mohanty, Prabhudutta, Manas Ranjan Kabat, and Manoj Kumar
7th ACM Conference on Embedded Networked Sensor Systems, Patel. "An Energy Aware Transport Protocol in Wireless Sensor
2009, pp 183-196. Networks." International Journal of Information (2012).
[8] http://www.libelium.com/smart_roads_wsn_smart_infrastructures. [33] Ahmed, Nadeem, Salil S. Kanhere, and Sanjay Jha. "The holes
[9] Imad Jawhar et.al. A framework for pipeline infrastructure moni- problem in wireless sensor networks: a survey." ACM SIGMO-
toring using wireless sensor networks, Wireless Telecommunica- BILE Mobile Computing and Communications Review 9.2 (2005):
tions Symposium, 2007. WTS 2007, pp. 1-7. 4-18.
[10] Alam Mainwarg et.al. Wireless Sensor Networks for Habitat [34] Horst F. Wedde and Muddassar Farooq. A Comprehensive review
Monitoring, WSA 02, Sept. 28, 2002, Atlanta, USA. of nature inspired routing algorithms for fixed telecommunication
[11] Flammini, F.; Gaglione, A.; Ottello, F.; Pappalardo, A.; Pragliola, networks. Journal of Systems Architecture, Vol. 52, 2006, pp. 461-
C.; Tedesco, A.; "Towards Wireless Sensor Networks for railway 484.
infrastructure monitoring," Electrical Systems for Aircraft, Rail- [35] Frederick Ducatelle, Gianni Di Caro and Luca M. Gambardella.
way and Ship Propulsion (ESARS), 2010, vol., no., pp.1-6, 19-21 AntHocNet: An Adaptive Nature-Inspired Algorithm for Routing
Oct. 2010. in Mobile Ad Hoc Networks. European Trans. On Telecommuni-
[12] Kameswari Chebrolu et al. BriMon: A Sensor Network System for cations, Vol. 16, 2005, pp. 443-455.
Railway Bridge Monitoring, Proceedings of the 6th international [36] Pruet Boonma, P. Champrasert, et al. BiSNET: A Biologically-
conference on Mobile systems, applications, and services, pp 2-14. Inspired Architecture for Wireless Sensor Networks.
[13] Flouri, K., et al.; (2012) A versatile software architecture for civil http://ieeexplore.ieee.org/stamp/stamp.jsp? a.r.number =01690264.
structure monitoring with wireless sensor networks. Smart Struc- [37] H. Ren and Max Q.-H. Meng. Biologically Inspired Approach for
tures and Systems, Vol. 10(3) 2012, pp 209-228. Wireless Sensor Networks. Proc. IEEE Int. Conf. Mechatronics
[14] Chen, Zhiliang, et al. "Two-way relaying for multiple applications and Automation 2006, pp. 762-768.
in wireless sensor networks." Smart Antennas (WSA), 2012 Inter- [38] Urgaonkar, Rahul, and Bhaskar Krishnamachari. "FLOW: An
national ITG Workshop on. IEEE, 2012. efficient forwarding scheme to mobile sink in wireless sensor net-
[15] Heidemann, John, et al. "Underwater sensor networks: applica- works." Proceedings of ACM SECON (2004).
tions, advances and challenges." Philosophical Transactions of the [39] Luo, Jun, et al. "Mobiroute: Routing towards a mobile sink for
Royal Society A: Mathematical, Physical and Engineering Sci- improving lifetime in sensor networks." Distributed Computing in
ences 370.1958 (2012): 158-175. Sensor Systems (2006): 480-497.
[16] Akyildiz, Ian F., Vuran Mehmet Can, Wireless Sensor Networks, [40] Song, Liang, and Dimitrios Hatzinakos. "Architecture of wireless
John Wiley and Sons, ISBN 978-0-470-03601-3, New York, 2010. sensor networks with mobile sinks: Sparsely deployed sensors."
[17] Ganesan et.al., Networking issues in wireless sensor networks, J. Vehicular Technology, IEEE Transactions on 56.4 (2007): 1826-
Parallel Distrib. Comput. 64 (2004) 799–814. 1836.
[18] Gowrishankar.S et al. Issues in Wireless Sensor Networks, Pro- [41] Li, Xu, Amiya Nayak, and Ivan Stojmenovic. "Sink mobility in
ceedings of the World Congress on Engineering 2008 Vol I. wireless sensor networks." Wireless Sensor and Actuator Networks
[19] Rajagopalan, R.; Varshney, P.K.; "Data-aggregation techniques in (2010): 153.
sensor networks: a survey," Communications Surveys & Tutorials, [42] Yang, Tao, et al. "Impact of Mobile Sink for Wireless Sensor
IEEE, vol.8, no.4, pp.48-63, Fourth Quarter 2006. Networks considering Different Radio Models and Performance
[20] Zia, Tanveer, and Albert Zomaya. "Security issues in wireless Metrics." Broadband, Wireless Computing, Communication and
sensor networks." Systems and Networks Communications, 2006. Applications (BWCCA), 2010 International Conference on. IEEE,
ICSNC'06. International Conference on. IEEE, 2006. 2010.
[21] Pathan, Al-Sakib Khan, Hyung-Woo Lee, and Choong Seon Hong. [43] Jordan, Edward, Jinsuk Baek, and Wood Kanampiu. "Impact of
"Security in wireless sensor networks: issues and challenges." Ad- mobile sink for wireless sensor network." Proceedings of the 49th
vanced Communication Technology, 2006. ICACT 2006. The 8th Annual Southeast Regional Conference. ACM, 2011.
International Conference. Vol. 2. IEEE, 2006. [44] Liu, Wang, et al. "Performance Analysis of Wireless Sensor Net-
works With Mobile Sinks." Vehicular Technology, IEEE Transac-
tions on 61.6 (2012): 2777-2788.
 36 International Journal of Sensors, Wireless Communications and Control, 2013, Vol. 3, No. 1 Gupta and Kumar

[45] Jerew, Oday, Kim Blackmore, and Weifa Liang. "Mobile Base Computer Communications and Networks, 2007. ICCCN 2007.
Station and Clustering to Maximize Network Lifetime in Wireless Proceedings of 16th International Conference on, vol., no.,
Sensor Networks." Journal of Electrical and Computer Engineering pp.1297,1302, 13-16 Aug. 2007 doi: 10.1109/ICCCN.2007.4318000.
2012 (2012). [53] Wang, Guojun, et al. "Adaptive location updates for mobile sinks
[46] Di Francesco, Mario, Sajal K. Das, and Giuseppe Anastasi. "Data in wireless sensor networks." The Journal of Supercomputing 47.2
collection in wireless sensor networks with mobile elements: A (2009): 127-145.
survey." ACM Transactions on Sensor Networks (TOSN) 8.1 [54] Prathap, U., et al. "Wireless Sensor Networks Applications and
(2011): 7. Routing Protocols: Survey and Research Challenges." Cloud and
[47] Majid I. Khan, Wilfried N. Gansterer, Guenter Haring “Static vs. Services Computing (ISCOS), 2012 International Symposium on.
mobile sink: The influence of basic parameters on energy effi- IEEE, 2012.
ciency in wireless sensor networks” Computer Communications, [55] Sastry, Anitha S., Shazia Sulthana, and S. Vagdevi. "Security
Available online 7 November 2012 http://dx.doi.org/10.1016/ Threats in Wireless Sensor Networks in Each Layer." Int. J. Ad-
j.comcom.2012.10.010 vanced Networking and Applications 4.04 (2013): 1657-1661.
[48] Li, Xu, Amiya Nayak, and Ivan Stojmenovic. "Sink mobility in [56] Koushanfar, Farinaz, Miodrag Potkonjak, and A. Sangiovanni-
wireless sensor networks." Wireless Sensor and Actuator Networks Vincentell. "Fault tolerance techniques for wireless ad hoc sensor
(2010): 153. networks." Sensors, 2002. Proceedings of IEEE. Vol. 2. IEEE,
[49] Rasheed, “Security Schemes for Wireless Sensor Network with 2002.
Mobile Sink” Ph.D Thesis, http://repository.tamu.edu/bitstream/ [57] Abedi, Raza H., Nauman Aslam, and Sayeed Ghani. "Fault toler-
handle/1969.1/ETD-TAMU-2010-05-7844/rasheed-dissertation.pdf? ance analysis of heterogeneous wireless sensor network." Electri-
sequence=3. cal and Computer Engineering (CCECE), 2011 24th Canadian
[50] Liang, Weifa, Jun Luo, and Xu Xu. "Prolonging network lifetime Conference on. IEEE, 2011.
via a controlled mobile sink in wireless sensor networks." Global [58] Branch, Joel, et al. "In-network outlier detection in wireless sensor
Telecommunications Conference (GLOBECOM 2010), 2010 networks." Distributed Computing Systems, 2006. ICDCS 2006.
IEEE. IEEE, 2010. 26th IEEE International Conference on. IEEE, 2006.
[51] Gyudong Shim; Daeyeon Park, "Locators of mobile sinks for wire- [59] de Souza, Luciana Moreira Sá, Harald Vogt, and Michael Beigl.
less sensor networks," Parallel Processing Workshops, 2006. ICPP "A survey on fault tolerance in wireless sensor networks." Interner
2006 Workshops. 2006 International Conference on , vol., no., Bericht. Fakultät für Informatik, Universität Karlsruhe (2007).
pp.6 pp.,164, 0-0 0, doi: 10.1109/ICPPW.2006.43 [60] Porter, John H., Paul C. Hanson, and Chau-Chin Lin. "Staying
[52] Xiaobing Wu; Guihai Chen, "Dual-Sink: Using Mobile and Static afloat in the sensor data deluge." Trends in ecology & evolution
Sinks for Lifetime Improvement in Wireless Sensor Networks," 27.2 (2012): 121-129.

Received: February 12, 2013 Revised: April 13, 2013 Accepted: April 26, 2013

View publication stats