An Improved DEEHC to Extend Lifetime of WSN

@ IJTSRD | Available Online @ www.ijtsrd.com
ISSN No: 2456
International
Research
An Improved DEEHC
Pragati Dikshit1
CSE Department
ABSTRACT
Sensor networks consist of randomly deployed nodes
that are powered by smaller batteries. The quantum
and quality of information sensed by these nodes is of
vital importance. Battery saving techniques or
algorithms are need of the hour. One such technique is
clustering of the nodes. This pap
modification to existing DEEHC algorithm that builds
multiple vertex disjoint paths from nodes to base
station via cluster heads. The proposed modification
(MDEEHC) improves the performance of the network
as equated to DEEHC and FLDEEC by relay
data of the child cluster heads using relay nodes.
MDEEHC shows improvement in terms of
throughput, packet delivery ratio, number of packet
drops and remaining energy.
Keywords: Sensors, disjoint paths, throughput,
clustering, MDEEHC, FLDEEHC, DEEHC
I. INTRODUCTION
The most challenging aspect of WSN is that they are
energy resource-constrained and that energy cannot be
replenished. The problem arises when all the sensor
nodes are required to forward the data packets to the
sink node. In this process, the available energy in each
node can be wasted through idle listening and
retransmitting due to collisions as well as overhearing.
Cluster-based WSN routing protocols excel network
topology management and energy minimization.
Clustering methods in WSN lead the sensor nodes to
be organized into small disjoint groups, where each
cluster has a coordinator referred as CH. In cluster
based approaches the sensors do not need to
communicate directly with BS. Instead, the CHs are
responsible to organize cluster members (CMs) and
send the data collected within the cluster to the BS.
@ IJTSRD | Available Online @ www.ijtsrd.com | Volume – 2 | Issue – 5 | Jul-Aug 2018
ISSN No: 2456 - 6470 | www.ijtsrd.com | Volume
International Journal of Trend in Scientific
Research and Development (IJTSRD)
International Open Access Journal
An Improved DEEHC to Extend Lifetime of WSN
1
, Rajveer Kaur Hyher2
, Goldendeep Kaur
1
Student, 2
Professor
Department, Guru Nanak Dev Engineering College,
Ludhiana Punjab, India
deployed nodes
that are powered by smaller batteries. The quantum
and quality of information sensed by these nodes is of
vital importance. Battery saving techniques or
algorithms are need of the hour. One such technique is
clustering of the nodes. This paper presents
modification to existing DEEHC algorithm that builds
multiple vertex disjoint paths from nodes to base
station via cluster heads. The proposed modification
(MDEEHC) improves the performance of the network
as equated to DEEHC and FLDEEC by relaying the
data of the child cluster heads using relay nodes.
MDEEHC shows improvement in terms of
throughput, packet delivery ratio, number of packet
Sensors, disjoint paths, throughput,
DEEHC
The most challenging aspect of WSN is that they are
constrained and that energy cannot be
replenished. The problem arises when all the sensor
nodes are required to forward the data packets to the
ocess, the available energy in each
node can be wasted through idle listening and
retransmitting due to collisions as well as overhearing.
based WSN routing protocols excel network
topology management and energy minimization.
WSN lead the sensor nodes to
be organized into small disjoint groups, where each
cluster has a coordinator referred as CH. In cluster
based approaches the sensors do not need to
communicate directly with BS. Instead, the CHs are
ster members (CMs) and
send the data collected within the cluster to the BS.
This process lead to a significant reduction in the
amount of transferred data in the network.
Consequently, overheads in communication as well as
energy consumption in cluster
significantly.
Fig.1 WSN Architecture
Maintaining the created clusters is the main
challenging task in the methods. To choose a node as
a CH, it is necessary to define its eligibility. That is
calculated based on local information of the nodes’
current situations many of the clustering techniq
talk about CH advertisement process in which every
eligible node will broadcast an advertisement message
for the election of CH. This process will consume
large amount of energy, as so many nodes will be
involved in the broadcasting process. In additio
multi-hop communication approaches followed
requires parent cluster heads to forward the data of
child CH as well. This increases load over the cluster
heads. This paper proposes modification to existing
DEEHC by adding the concept of choosing optimal
Aug 2018 Page: 676
6470 | www.ijtsrd.com | Volume - 2 | Issue – 5
Scientific
(IJTSRD)
International Open Access Journal
f WSN
Goldendeep Kaur2
This process lead to a significant reduction in the
ransferred data in the network.
Consequently, overheads in communication as well as
n in clustering will be reduced
Fig.1 WSN Architecture
Maintaining the created clusters is the main
challenging task in the methods. To choose a node as
a CH, it is necessary to define its eligibility. That is
calculated based on local information of the nodes’
of the clustering techniques
talk about CH advertisement process in which every
eligible node will broadcast an advertisement message
for the election of CH. This process will consume
large amount of energy, as so many nodes will be
involved in the broadcasting process. In addition,
hop communication approaches followed
requires parent cluster heads to forward the data of
child CH as well. This increases load over the cluster
heads. This paper proposes modification to existing
DEEHC by adding the concept of choosing optimal

International Journal of Trend in Scientific Research and Development (IJTSRD) ISSN: 2456-6470
@ IJTSRD | Available Online @ www.ijtsrd.com | Volume – 2 | Issue – 5 | Jul-Aug 2018 Page: 677
relay nodes to forward data from child cluster heads
to base station. The remainder section of this paper is
organized as follows: Section II of the paper surveys
the related work, Section III depicts the research
methodology, and Section IV portrays the outcomes
we got from the work. Section V describes the
conclusion and future work.
II. Literature Review
Numerous plans have just been proposed with respect
to adaptation to non-critical failure in WSN to
accomplish unwavering quality, unpredictability,
vitality sparing and so forth. There are a portion of the
papers discussed here which have talked about
adaptation to internal failure methods.
Prasenjit Chanak et. al, [1], reflected the difficulty
of enduring system failures realized by positioned
sensors in a sensor network. They initially suggested a
new clustering procedure for sensor networks, named
DEEHC that chooses CHs conferring to the remaining
energy of positioned nodes with the help of a
subordinate timer. For the duration of the clustering
stage, individual node discovers k-vertex separate
routes to CHs reliant on the energy level of its
neighbor nodes. They also suggest a kVDPR
procedure where respective CH discovers k-vertex
separate routes to the BS and communicates their
amassed information to the BS. Additionally, they
furthermore suggest a new RMM that can restore k-
vertex separate routes through the observing period.
The subsequent network turns out to be tolerant to k-1
disasters in the nastiest situation. The suggested
system has been comprehensively verified by means
of numerous network situations and associated to the
prevailing methodologies to demonstrate its
efficiency.
Sushruta Mishra et. al., [2] tried to cover issues
identified with fault recognition and recovery systems
in wireless sensor networks in both hypothetical and
application situated research. Reliability was an
essential issue in a network and when the network is
energy constrained as in a wireless network then it
turns into a high need to guarantee that this network is
profoundly dependable. Be that as it may, this
reliability is seriously influenced by mistakes and
faults that happen because of different reasons, for
example, Hardware failing, bugs in programming,
natural dangers and so forth. Accordingly, a sensor
network ought to be fault safe so it can manage these
incorrect conditions adequately.
Prasenjit et. al., [3] proposed an energy productive
node fault determination and recovery for wireless
sensor networks alluded as fault tolerant multipath
routing plan for energy effective wireless sensor
network (FTMRS). The FTMRS depends on
multipath information routing plan. One most limited
way is use for principle information routing in
FTMRS method and other two backup ways are
utilized as option way for faulty network and to deal
with the over-burden activity on primary channel.
Most limited way information routing guarantees
energy productive information routing. The execution
examination of FTMRS shows better outcomes
contrasted with other well-known fault tolerant
methods in wireless sensor networks.
Samira et. al., [4] proposed characterization in view
of the network measure, since the execution of the
larger part systems relies on upon the size regarding
geographical area and number of nodes. Fault
tolerance is a standout amongst the most vital wireless
sensor networks requirements. It guarantees that the
network keeps on working accurately even when a
few components fail. Truth be told, fault tolerance is a
need in this sort of networks because of sensor node
attributes, radio correspondences and threatening
conditions in which these networks are deployed. In
this study, they give a diagram of WSN components
that give or enhance the fault tolerance property of
wireless sensor networks.
Savita et. al., [5] displayed a basic investigation of
different fault tolerance components in wireless
sensor networks, for example, excess based
instruments, clustering based systems and
arrangement based systems to recognize the qualities
and shortcomings of every one of these components.
At long last, the paper presents conclusion and
recommends some future research headings that will
be useful for specialists who are working in this field.
Fault tolerance is one of the basic issues in Wireless
Sensor Network (WSN) applications. The issue of
missing sensor node, correspondence connection and
information are inescapable in wireless sensor
networks. WSNs encounter disappointment issues
because of different components, for example, control
exhaustion, ecological effect, radio obstruction,
disengagement of sensor node and impact. Numerous
specialists have proposed fault tolerant instruments
that can accomplish higher information reliability,
precision, energy saving, improve network lifetime

and limit disappointment of segments of wireless
sensor network.
Doo soon park [6] proposed a measure to distinguish
and manage sensors by partitioning WSN into settled
WSN and portable WSN. This paper additionally
proposes a strategy to predict and express the energy
utilization of a sensor network. The
commercialization of recognition and observation
frameworks by sensor networks has as of late turned
out to be more prominent in different ventures and in
common life. Wireless sensor networks (WSNs)
might be used in territories. WSNs, which have three
essential components, ought to have the capacity to
convey through a blend of network and sensor
capacities. In expansive scale sensor networks,
however, support expenses to replace nodes when
they are out of request, and batteries utilized for
wireless interchanges, extraordinarily increment,
notwithstanding the cost of building them. In this
way, measures to distinguish and react to sensor faults
in a sensor network, and additionally an instrument to
predict the energy utilization of sensors, are essential.
Cho H., et. al., [7] proposed Construction algorithms
for reliable CDS (CARCODS). This is the technique
by which the execution of specially appointed
topologies that comprise of CDSs is made strides. A
CDS arrangement technique was recommended by
making utilization of the telecom time deferral of the
neighboring setup notice message that components in
the rest of the battery limit, the portability of nodes,
and the quantity of neighboring nodes.
Hong Y. et. al., [8] proposed partial reconstruction of
CDS (PRCDS). A CDS fractional reconfiguration
algorithm was proposed to adequately respond to
issues because of the event of basic nodes when a
CDS-based routing convention was connected. In
designing the CDS nodes for load adjusting, this
strategy scans for the likelihood of association, while
topology reconfiguration, because of basic nodes, is
expanded into two-hope nodes. Subsequently, the
nodes inside a particular range in terms of their
reconfiguration time.
Zhang, et, al [9] proposed a fuzzy Logic based
distributed energy-efficient clustering algorithm
(FLDEEC). During cluster head (CH) election, it
takes nodes’ energy, nodes’ degree and neighbor
nodes’ residual energy into consideration as input
parameters. Through fuzzy interference system, node
uses its own information and its neighbor nodes’
information to make a judgment to decide whether it
fits to be a CH or not in a distributed way. FLDEEC
algorithm can solve hot spot problem caused by
communication in cluster, improve the efficiency of
data transmission and extend the lifetime of the whole
network by screening out nodes with bad conditions.
The experimental results indicated that the FLDEEC
algorithm is better than CHEF algorithm and DEEC
algorithm in the aspects of data transmission, energy
consumption and the number of living nodes.
III. Research Methodology:
The proposed scheme works in the following manner:
Modified DEEHC algorithm
Initially each sensor node will check its current
energy level. If the current energy level is more than
the threshold value, then it starts advertisement
process where it broadcasts advertisement message to
all the sensors in its communication range. The
message contains its ID, location information and its
residual energy. During this process, if any other
node, which has received the advertisement packet,
checks that its current energy is less than received
energy value in the advertisement packet then it
withdraws itself from advertisement process and
becomes cluster member. If any node receives
multiple advertisement messages, then it would join
the cluster head for which energy is highest. When the
non-CH node broadcasts a join packet to the CH to
join its cluster, it would additionally broadcast its
location information and residual energy as well. The
current CH would therefore collect all the data from
its cluster members. This data will be used to rotate
the CHs for the subsequent rounds. The current CH
will sort the members in the order of highest energy
and minimum distance from the base station. The
member at the top of the list would be selected as CH
for the subsequent round. Thus, at the beginning of
every round, the CH advertisement phase would be
avoided.
Modified k-vertex disjoint paths routing
Once the clusters have been formed and nodes have
been arranged in the order of becoming CHs for the
subsequent rounds, the base station will begin
broadcasting of L-ADVE to the cluster heads in its
communication range. The output of the existing
scheme would be multiple disjoint paths from CHs to
the BS. Once the various paths have been formulated,
the every CH would select a member node from its
cluster to relay the data of the child CH. The member
node would be the one having highest residual energy

and least distance from the midpoint of child CH and
BS. Thus, every CH would now relay the data to the
chosen relay node which would act as parent to child
CH. This would reduce the load over the CH to relay
the data of other CHs as well.
Modified Route Maintenance Mechanism
In this mechanism, the focus will be given on the
disjoint path made between the CHs and BS via
chosen relay nodes. If residual energy of any chosen
relay node falls to less than 80 percent of the initial
energy value or any node goes faulty then the next
cluster member will act as relay node for the child
CH. This would further lead to better routes formed in
the network.
Algorithm of the modified DEEHC
1. Suppose ‘N’ is the number of the nodes in the
network
2. For i=1:N
A. If Remaining Energy(i) > Threshold
I. Broadcast ADV packet in the
communication range
B. Else
I. Do not opt for broadcasting
C. End if
D. If Remaining energy < Energy received in
the ADV packet
I. Withdraw from advertisement
E. Else
I. Continue with broadcasting process
F. End if
3. End for
4. Suppose ‘M’ be the number of nodes which
received ADV packet
5. For i=1:M
A. If ADV packet is received from multiple
CHs
I. Form k-vertex paths with them
II. Choose highest energy CH as
parent CH
B. Else
I. Keep cluster head in the parent CH
list
C. End if
6. Form cluster with the selected parent CH
7. Send Join packet to the CH
8. End for
9. For each CH, do
A. Arrange members in increasing order of
residual energy and least distance from BS
B. Choose node at top of the list to be CH for
the next round
10. End for
11. Base station broadcasts L-ADVE packets to CH
12. Each CH re-broadcasts L-ADVE packet to child
CHs
13. Suppose ‘k’ is the number of disjoint paths from
CH to BS
14. For i=1:k
A. Each CH chooses relay node from its
member nodes
15. End for
16. If any node goes fault
A. CH chooses next member as relay node for
child CH
End if
IV. Implementation & Results
In this work, three techniques have been simulated
namely FLDEEC [9], DEEHC [1] and third was
modification to DEEHC, which we named Modified-
DEEHC or MDEEHC. All the three techniques are
simulated in Network Simulator 2.35. The following
set of input parameters were considered while
creating a wireless sensor network:
Parameter Value
Channel Wireless
Mac 802.11
Propagation Model Two Ray Ground
Antenna Omni Directional
Number of nodes 100
Network Area 700 * 750 sq meters
Initial Energy 20 Joules
Queue CMUPriqueue
Queue Length 500
Table 4.1: Simulation Parameters
The performance of the network was measured based
on throughput, packet delivery ratio, number of
packet drops and energy remaining in the network.

Figure 4.1: Remaining energy
The initial average energy of the network was 20
Joules. At the end of simulation, the remaining energy
was approx. 9 Joules for M-DEEHC, 6.5 Joules for
DEEHC and 2 Joules for FLDEEC. The amount of
remaining energy is highest for proposed technique.
Initially, all the three schemes opt for cluster
formation process. This process requires substantial
amount of broadcasting leading to consumption of
energy sources of the sensors. This process is run
every time at the start of new round for the cluster
head selection process. The proposed scheme,
however, avoids successive runs of this process by
electing the cluster heads for the subsequent rounds in
the current round itself. Thus, nodes do not need to
advertise themselves for the cluster head election
process. The proposed scheme proposes that cluster
heads do not take on the load of forwarding the data
of the child cluster heads. Instead, the parent cluster
heads choose a relay node from their clusters to
forward the data from child cluster head to the base
station. Thus, these two reasons when combined,
leads to lesser energy consumption of the network as
well as increased lifetime.
Figure 4.2: Packet Delivery Ratio
In M-DEEHC, the cluster heads for the subsequent
rounds are selected in the previous round which
avoids broadcasting of ADV packets in consecutive
rounds. Thus, the least amount of broadcasting results
in least congestion and best value for packet delivery
ratio for M-DEEHC. Consequently, after the cluster
formation, value of PDR increases during data packet
transmission in the network. The value for PDR is
0.91 for M-DEEHC, 0.76 for DEEHC and 0.72 for
FLDEEC.
Figure 4.3 Number of Packet Drops

It can be seen from the graph that our proposed
scheme shows very less number of packet drops in the
network. This is because of reduced congestion (by
prevention of multiple times broadcasting of ADV
packets) in the network. The number of packet drops
were recorded highest for FLDEEC scheme and lesser
for DEEHC.
Figure 4.4: Throughput
This indicates amount of data received at the
destination node per unit of time. Lesser packet drops
leads to increase the throughput of the network for M-
DEEHC. The value of throughput is in line with the
value obtained for PDR as well. For M-DEEHC, the
value of throughput was around 8500 Kbps, for
DEECH this value was around 6800 Kbps and for
FLDEEC the value for 5700 Kbps.
V. Conclusion and Future Work
In this work, we have implemented FLDEEC,
DEEHC and modification to existing DEEHC
clustering algorithm. FLDEEC defines the selection
of cluster head according to fuzzy rules. DEEHC
defines the formation of k-vertex disjoint paths from
cluster members to cluster heads and from cluster
heads to sink node or base station. This algorithm
selects cluster heads considering remaining energy of
the nodes as prime factor. M-DEEHC modifies the
cluster formation of DEEHC algorithm as well as
modifies the way the data is transmitted from child
cluster heads to the base station. The performance of
the network was analyzed based on throughput,
remaining energy, number of packet drops and packet
delivery ratio. The initial average energy of the
network was 20 Joules. At the end of simulation, the
remaining energy was approx. 9 Joules for M-
DEEHC, 6.5 Joules for DEEHC and 2 Joules for
FLDEEC. The amount of remaining energy is highest
for proposed technique. In M-DEEHC, the cluster
heads for the subsequent rounds are selected in the
previous round, which avoids broadcasting of ADV
packets in consecutive rounds. Thus, the least amount
of broadcasting results in best value for packet
delivery ratio for M-DEEHC. The value for PDR is
0.91 for M-DEEHC, 0.76 for DEEHC and 0.72 for
FLDEEC. This also leads to lesser values for number
of packet drops in the network. For M-DEEHC, the
value of throughput was around 8500 Kbps, for
DEECH this value was around 6800 Kbps and for
FLDEEC the value for 5700 Kbps. Thus, it can be
concluded that our proposed scheme outperforms the
other schemes.
In future, other parameters such as delay, jitter etc.
can be analyzed for the proposed scheme. These days,
the concept of Internet of Things is in high demand.
This work can also be explored for an application of
IoT such as military surveillance, forest fire detection
etc. Internet of Things combine sensor networks along
with other networks such as MANETs, VANETs or
FANETs. M-DEEHC can be explored for the same.
References
1. Prasenjit Chanak, Indrajit Banerjee, R. Simon
Sherratt, “Energy-Aware Distributed Routing
Algorithm to Tolerate Network Failure in
Wireless Sensor Networks”, Ad Hoc Networks,
2016.
2. Sushruta Mishra, Lambodar Jena and Aarti
Pradhan, " Fault Tolerance in Wireless Sensor
Networks" in International Journal of Advanced
Research in Computer Science and Software
Engineering Oct 2012.
3. Prasenjit Chanak, Tuhina Samanta, Indrajit
Banerjee, "Fault-Tolerant Multipath Routing
Scheme for Energy Efficient Wireless Sensor
Networks" in international Journal of Wireless &
Mobile Networks (IJWMN) Vol. 5, No. 2, April
2013.
4. Samira Choukhi," A survey on fault tolerance in
small and large scale wireless sensor networks" in
ELSEVIER Sep 2015.

5. Savita Gupta and Sukhwinder singh, “A survey
on fault tolerance techniques in Wireless Sensor
Networks" in IEEE 2016.
6. Doo soon park, "Fault Tolerance and Energy
Consumption Scheme of a Wireless Sensor
Network" in international journal of distributed
sensor Nov 2013.
7. Cho H. -S., Yoo S.-J, “Power, mobility and
wireless channel condition aware connected
dominating set construction algorithm in the
wireless ad-hoc network "in the Journal of Korea
Information and Communications Society 2005.
8. Hong Y. -S., Lim H. -S. A, “load-balanced
topology maintenance with partial topology
reconstruction" in the Journal of Korea
Information and Communications Society 2010.
9. Ying Zhang, Rundong Zhou, “Networks
Energy-Efficient clustering algorithm based on
fuzzy interference system”, 29th Chinese Control
and decision policies (CCDC), IEEE, 2017.

An Improved DEEHC to Extend Lifetime of WSN

More Related Content

What's hot (20)

Similar to An Improved DEEHC to Extend Lifetime of WSN (20)

More from ijtsrd (20)

Recently uploaded (20)

An Improved DEEHC to Extend Lifetime of WSN