Optik 127 (2016) 6590–6600

Contents lists available at ScienceDirect

Optik
journal homepage: www.elsevier.de/ijleo

Original research article

A survey on LEACH and other’s routing protocols in wireless

sensor network
Vishal Kumar Arora (Research Scholar) a , Vishal Sharma b,∗ , Monika Sachdeva a
a
Department of Computer Science & Engineering, Shaheed Bhagat Singh State Technical Campus, Ferozepur, Punjab 152004, India
b
Department of Electronics & Communication Engineering, Shaheed Bhagat Singh State Technical Campus, Ferozepur, Punjab 152004,
India

a r t i c l e i n f o a b s t r a c t

Article history: An extensive range of applications of Wireless Sensor Networks such as military, envi-
Received 17 February 2016 ronment, surveillance, home, vehicle tracking/detection, traffic flow and medical make it
Accepted 13 April 2016 hot-spot in the epoch of wireless networks. A WSN consisting of numerous sensor-nodes
is equipped with inadequate energy, memory, and computation capability issues. Further,
Keywords: such networks are limited to reinstate the dead nodes caused by energy’s depletion and
Clustering to maximize the life-span of the system. To achieve this aim, several routing algorithms
LEACH protocol
are proposed and investigated. In this work, an attempt is carried out to assess the diverse
Network routing
hierarchical routing protocols, developed from LEACH and is extended to other presented
Energy
Wireless sensor network routing protocols like TEEN, APTEEN, and PEGASIS. Depending upon the observations and
scrupulous consideration, a relative conclusion is drawn in the last.
© 2016 Elsevier GmbH. All rights reserved.

1. Introduction

Wireless Senor Network (WSN) is a collection of large number of small size and moderately inexpensive computational
nodes that forward the valuable information to a central point for appropriate processing. The environment can be an
information technology framework, a biological system or a physical work. There are four parts of sensor network: (i)
sensors (ii) network connecting different sensors (iii) centralized information gathering store (iv) resources performing
computation which include data mining, data correlation etc. [1–3]. Sensors nodes make an ad hoc network that are useful
to monitoring temperature, pressure, humidity, military surveillance, disaster management, forest fire-tracking and many
more [4]. Routing in WSN is different from other wireless network due to sensor node have constraints of energy, processing
activities, transmitting collected data from multiple nodes to a single sink, unique global address is not possible due to
random deployment of nodes etc. Due to all of these reasons, different types of routing protocols were developed for such
scenarios. All these routing protocols had considered all those inherited features of WSNs. Main aim of these protocols were
to reduce power consumptions and increasing network life time. This can be achieved by implementing routing protocols
that consume minimum energy, choose path between sensor nodes and base station in such manner that increase network
life time. Basically, WSN routing protocols are classified into four main categories: Network structure, communication
models, topology based and reliable routing schemes [5]. Network structure protocols, basically, rely upon the architecture

∗ Corresponding author.
E-mail addresses: [email protected] (V.K. Arora (Research Scholar)), er [email protected], [email protected] (V. Sharma),
[email protected] (M. Sachdeva).

http://dx.doi.org/10.1016/j.ijleo.2016.04.041
0030-4026/© 2016 Elsevier GmbH. All rights reserved.
 V.K. Arora (Research Scholar) et al. / Optik 127 (2016) 6590–6600 6591

Fig. 1. LEACH clustering hierarchical model [7].

of network. Routing protocols in this category are differentiated on basis of nodes interconnection and route they follow to
transmit data packets from source to destination. This leads to following types of categorisation as:

• Flat Protocol: Nodes are deployed uniformly and have same role i.e. every node is at same level inside network. FLAT
protocols can be further classified as; pro-active, reactive and hybrid protocols [6].
• Hierarchical Protocols: In these types of protocols nodes are arranged into clusters and node having maximum energy
becomes CH (cluster-head) of the cluster. Cluster-head coordinates activities inside and outside cluster. Cluster-head is
responsible for collecting data from nodes of their cluster and removing redundancy among collected data to reduce energy
requirement for transmitting of data packets from cluster-head to base station e.g. LEACH, SEP, TEEN, APTEEN etc. [7,8].
• Location based Protocols: Nodes are distinguished on basis of their location inside network. Distance among sensors nodes
are calculated on basis of signal strength, higher the signal strength lesser the distance between nodes. Some protocols in
this category allow nodes to go into sleep mode if there is no activity going on at that node e.g. GEAR and GPSR [9,10].

Among these categories of routing protocols of WSN, FLAT protocols have minimum overhead to maintain resources
among communicating nodes [6] and hierarchy protocols reduces the size of routing tables [8]. Main aim of this paper is to
review the hierarchical energy efficient routing protocols along with modifications over some of these protocols.

2. Hierarchaical protocols

In hierarchical routing protocols, nodes organize the network into set of clusters. Each cluster is managed by a selected
cluster-head [11]. Cluster-head periodically collects data from member nodes of their cluster, compresses it and then removes
duplicacy among collected data to reduce the number of transmission between cluster-head and base-station.

2.1. LEACH (Low energy adaptive clustering hierarchy) and its variants

LEACH is a routing protocol that collects and sends data to base station with following main objectives [7]:

• Increase network life-time.

• Decrease energy dissipation of sensor nodes.
• Reduce the number of communication messages.

To attain these objectives, nodes organizes themselves into clusters. As shown in Fig. 1, member nodes of a cluster sends
their respective data to their cluster-head, which is further responsible for sending collected to base station. This results
in saving the energy of sensor node because they have to spend lesser energy to send their data to cluster-head instead of
base-station.
Moreover, cluster-heads aggregates collected data to remove redundancy among similar data and hence reduces the
transmitted data to the base station. This results in saving large amount of energy, as aggregated data is sent over a single
hop. LEACH operates in two diverse phases including setup- and state-phase. The setup phase is further categorised as
cluster-head selection and cluster formation. Cluster-head selection ensures that this role rotates among all sensor nodes;
to evenly distribute energy consumption among all network nodes. So, selected cluster-head last long only for a round and
this role is rotated among other nodes so that selected cluster-head did not die soon. To find out its turn to act as cluster-
head, node ‘n’ generates a random number between 0 and 1 and compare with the cluster-head selection threshold T (n).
A node becomes cluster-head if its generated number is less than a threshold T (n) [4,11]. Cluster-head threshold ensures
two things: First, only predetermined fractions of nodes, P, become cluster-head. Second, node which acts as cluster-head
in last 1/P rounds are not selected as cluster-head. To meet these requirements threshold is set as:
 ∗
P/1-p (r mod1/p) if nD G
T(n) =
0 otherwise
6592 V.K. Arora (Research Scholar) et al. / Optik 127 (2016) 6590–6600

where, r is the current round, and G is the set of nodes that have not been cluster-heads in the last 1/p rounds. At the end
of cluster-head selection process, every node selected as cluster-head advertise its new role to rest of the network. After
receiving the cluster-head advertisement, remaining nodes selects a cluster to join on basis of received signal strength and
inform their selected cluster-head of their wish to become a member to this cluster. After cluster formation, each cluster-
head creates and distributes TDMA schedule among each member of their cluster. This ends the setup phase and starts the
setup state phase. During steady phase each node transmits sensed data to cluster-head during its allocated time slots.

3. Different LEACH variant’s

Non-uniform distribution of cluster-head is the main drawback of LEACH that leads to early fading of battery of cluster-
head and hence, lowers the network life-time. This limitation is overcome in descendants of LEACH protocol which are
discussed as:

3.1. C-LEACH (Centralized LEACH)

In LEACH, every node takes its own decision to become a cluster-head, but the main limitation is that there is no guarantee
for cluster-head position and number of nodes in each cluster. In other words, LEACH algorithm does not guarantee about
the location of cluster-head, which may effectively decrease the overall performance during some rounds.This may makes
LEACH ineffective protocol. C-LEACH, a central control algorithm is used to make clusters in such a way that cluster-heads
are scattered throughout the network [12]. This algorithm is implemented at base station, which selects nodes to make
them cluster-head for current round. Every sensor node sends their current location (using GPS) and residual energy to the
base station. Base station calculates the average node energy and nodes that have energy less than average cannot become
cluster-head for that round. After cluster-head formation, base station broadcast this information to all nodes in the network,
using a message which contains the cluster-head ID for every node. Node having same ID became cluster-head for that round
and rest of nodes goes into sleep mode until their turn comes up for data transmission. Cluster formed on such basis are
better than the cluster formed in LEACH. Steady-state phase of C-LEACH is same as that in LEACH, every node sends their
data to cluster-head and cluster-head after doing data aggregation sends compressed data to base station. Overall, C-LEACH
performs better than LEACH because it consider position of nodes inside network and creates cluster-head in such a way
that they are scattered throughout network which makes loads to distribute evenly among the cluster-heads.

3.2. MODLEACH

MODLEACH another cluster based algorithm differs from LEACH mainly on two points. One, there is no need to change
cluster-head until and unless it has more energy than the certain required threshold. Second, MODLEACH did not amplify
all the signals to same level.
In LEACH, cluster-head is changed after every round, so as to save the cluster-head to die early. But, in MODLEACH current
cluster-head is replaced by new one only if current cluster-head does not have energy less than the required threshold. It
saves energy consumed in cluster formation and forwarding the routing packets for searching another new cluster-head
[13]. In each round, if the residual energy of current cluster-head finds to be more than the minimum threshold value,
then the current cluster-head will remain cluster-head for new round. MODLEACH categorised communication into three
categories: (1) Intra cluster communication (2) Inter cluster communication (3) Data transmission from cluster-head to
base station. Energy required for intra cluster communication is different from inter cluster or cluster-head to base station
communication. So, different kind of amplification is required for different packets depending upon their type. Earlier in
LEACH all packets are being amplified in same manner irrespective of type of communication.

3.3. Stable election protocol (heterogeneous LEACH)

In LEACH, every sensor node is initialized to same energy level but in Stable Election Protocol (SEP) there are two different
types of nodes called as normal nodes and advanced nodes [14]. These nodes have different initial energy. There are m
numbers of advance nodes in network with ␣ additional energy. Advance node have energy Eo* (1 + ␣) where Eo is energy
of normal nodes [14]. Advance nodes are made cluster-heads more often as compared to normal nodes because advance
nodes have more energy as compared to normal nodes. So, in SEP initial energy is increased by ␣*m. and hence the overall
life time of network increases, so instability period decreases. The observations are evaluated by considering the following
parameters given in Table 1:
If death of node occurs early in the network, it results in the early death of other nodes in the network. As depicted in
Table 2, it is observed that the first node of SEP dies in the last and offers the highest stability period even after increasing
the node density. This is due to fact that in SEP advance nodes are made cluster-head more often and energy required for
cluster-head replacement is saved. Further, it is observed that MOD-LEACH outperforms as nodes density increases because
it saves energy that incurred on replacement of current cluster-head with new cluster-head.
 V.K. Arora (Research Scholar) et al. / Optik 127 (2016) 6590–6600 6593

Table 1
Performance Parameters.

Parameters Values

Sink Location 50,50

Network Size 100 m
Number of nodes 100
CH probability 5%
Initial node power 0.5 J
Nodes Distribution Nodes are uniformly distributed
Control Packet Size 50 bits/s
Data Packet size 1000 bits/s
Energy dissipation (Efs) 10 pJ/bit/m2
Energy for Transmission (ETX ) 50 nJ
Energy for Reception (ERX ) 50 nJ
Energy for Data Aggregation (EDA) 5 nJ/bit/signal

Table 2
Comparison of Network life-time (first node dead) for LEACH, CLEACH, MOD LEACH and SEP.

No. of nodes Round number when first node dies

LEACH Centralized LEACH MOD LEACH SEP

100 757 1012 941 1012

200 783 1023 1003 1038
300 790 997 1017 1098
400 785 950 1019 894
500 772 921 981 930

Table 3
Comparison of percentage of dead node for LEACH, CLEACH, MOD LEACH and SEP.

Percentage of Dead Nodes (%) LEACH Centralized LEACH MOD LEACH SEP

1 755 978 1001 1068

20 989 1074 1109 1204
50 1156 1156 1200 1298
70 1232 1212 1287 1361
90 1384 1334 1407 1460
100 1665 1677 1569 1832

Table 4
Comparison of first node and last node dead in LEACH, CLEACH, MOD LEACH and SEP.

Initial Energy (J/Node) LEACH Centralized LEACH MOD LEACH SEP

First node dies Last node dies First node dies Last node dies First node dies Last node dies First node dies Last node dies

0.25 365 841 461 795 448 778 512 1225

0.5 755 1665 978 1677 1001 1569 1068 1832
1 1507 3842 2028 2963 2116 3163 1884 3953

Table 3 shows that SEP again outperforms among the others and possesses maximum number of rounds in achieving
100% of dead nodes. This is because of making higher energy nodes as cluster-head, SEP saves energy to be incurred on
cluster- head replacement. Also, it is apparent from the observations, depicted in Table 3, that as percentage of dead nodes
increases, life-span of the network decreases. This happens because it creates more burdens on alive nodes that leads to
more energy consumption and hence, shorter life.
Table 4 shows comparison of LEACH, Centralized LEACH, MOD LEACH and SEP over different initial energy of node.
As it is clear that as the initial energy of node increases, the death of first node and last node gets delayed in all protocols.
It is also observed that in all of these protocols, the life-span of first- and last-node increases approximately two times on
increasing the initial energy of node by twofold. For initial energy of 0.25 J and 0.5 J, SEP shows better results but at 1 J of
initial energy, MOD-LEACH outperforms among the other protocols. This is occurred due to overhead replacement of current
cluster-head with new one is more and consumes the energy (Table 5).

3.4. MH- LEACH (Multi hop LEACH)

MH-LEACH protocol makes further development in LEACH to save energy by using nodes that lies on the way to base
stations [15]. MH-LEACH uses same practice of LEACH to select the cluster-heads and cluster formation. Role of cluster-
head is also same i.e. performing data fusion to the received packets so as to reduce the transmitting and forwarding
data in the network [4,11]. But, multi-hop LEACH possesses more life-span of nodes. This is because, during inter-cluster
6594 V.K. Arora (Research Scholar) et al. / Optik 127 (2016) 6590–6600

Table 5
Comparison of LEACH with its variants.

Protocol Difference from LEACH

Centralized-LEACH Base station is responsible for making cluster by considering current position
and remaining energy of sensor nodes. Base station implements optimal
algorithm to select cluster-head for current round
Modified LEACH Modified LEACH did not change cluster-head until and unless it has more
energy than the certain required threshold. Also, MODLEACH amplify signals
on basis of their packet types
Stable Election Protocol SEP has heterogeneous nodes called as normal nodes and advanced nodes.
Advance nodes have higher energy than normal nodes, so advance nodes are
made cluster-heads more often as compared to normal nodes
Multi hop LEACH Cluster-head instead of sending collected packet directly to base station uses
sensor nodes that lie on way to base station. Protocol increases network life
time by using neighbour nodes for data transmission which results in lesser
energy consumption
Two Level LEACH In TL LEACH there are two types of cluster-heads: secondary cluster-head and
primary cluster-head. TL LEACH divides task of collection and transmission to
secondary cluster-head and primary cluster-head respectively
Vice-LEACH In V- LEACH there is extra cluster-head called as vice cluster-head that
performs duty of cluster-head in case cluster-head dies

communication, cluster-head sends data packets to nearby cluster-head that lies on way to base station instead of sending
it directly to base station and hence saves energy of cluster-head. In addition, in intra-cluster communication, nodes instead
of sending data packets directly to cluster-head, it sends data to neighbour nodes that lies on way to cluster-head and saves
energy. Intermediate cluster-head or node makes a decision at its own depending upon their energy level whether to accept
data packets or not. So, if a cluster-head or node did not accept data packets, sensor node try to locate another cluster-head
or node as per its routing Table entry.
Multi hop LEACH protocol operates in two phases. In phase 1, all the cluster-heads broadcast an announcement message
and construct their routing Table depending upon level of signal (RSSI) received. Then, they make their route to base station
via closest cluster-head. In phase 2, each cluster-head sends these initial routes (from routing table) to the base-station. The
performance of MH-LEACH is investigated by designing a network of 200 m × 200 m consisting of 100 nodes with sink at a
distance of 50 m from the network [16]. It is reported that the designed network considered cluster-head probability of 10%,
packet size of 4000 bits, Efs of 10 pJ, ETX of 50 pJ, ERX of 50 pJ, and EDA of 5 pJ. Each node is fed by initial energy of 0.5J. The first
node of simulated MH-LEACH died after 70–80 rounds as compared to LEACH because MH-LEACH saves energy by instead
of transmitting data directly to base station uses intermediate node or cluster head for communication. Also, this is reason
behind the 25–30 percent more network life time of MH-LEACH as compare to LEACH for varying packet size (between 1000
and 10,000 bits).

3.5. TL-LEACH (Two levels LEACH)

TL-LEACH solves the uneven energy distribution problem of LEACH that occurs due to random selection of cluster-heads.
Earlier, cluster-heads were selected randomly, so different cluster-heads had different residual energy. If any low energy
node or node far away from base station becomes cluster-head, cluster-head dies quickly. TL-LEACH solves this problem
by dividing task of collection cum aggregation of data from nodes within cluster and transmission of collected data to base
station into secondary and primary cluster-heads respectively [17–19]. Secondary cluster-head is responsible for collecting
and aggregating data collected from member nodes and then forwarding to primary cluster-head; primary cluster-head is
responsible for transmitting received data from secondary cluster-head to base station. TL-LEACH adheres to cluster-head
selection and cluster formation process of LEACH protocol. TL LEACH evaluates following two conditions:

n 
• Current Cluster-head (Ecur ) energy less than the average energy (Eavg ) i.e, Ecur < Eavg whereEavg = E(i)cur .
n−1
• Distance (d) between cluster-head and base station is larger than the average distance (davg ) i.e, d > davg where davg =
nn=1 di .

If primary cluster-head has either lesser energy as compare to average energy or have distance more than average
distance, then another node is chosen having maximum energy in the cluster to act as secondary cluster-head. Secondary
cluster-heads creates and distributes TDMA schedule among their member nodes. If there did not exist any secondary node,
then it is responsibility of primary cluster-head to create and distribute TDMA schedule among the member nodes.
The author evaluated the performance of TL-LEACH, over a network of size 200 × 200 m having 200 nodes with each
node have an initial energy of 0.5J and sink at a distance of 100 m from the network [18]. The designed network considered
cluster-head probability of 7%, packet size of 4000 bits, Efs of 10 pJ, ETX of 50 pJ, ERX of 50 pJ, and EDA of 5 pJ. It has been
observed that the total energy of drained out after 4000 rounds as compared to 3000 rounds for LEACH in the network
 V.K. Arora (Research Scholar) et al. / Optik 127 (2016) 6590–6600 6595

Fig. 2. V-Leach Protocol [20].

Fig. 3. Chaining in PEGASIS.

consisting of 100 nodes placed randomly in area of 100 m × 100 m. The main reason behind the prolonged life-span of the
network incorporated TL-LEACH is due to the fewer cluster-head replacement.

3.6. V- LEACH (Vice-cluster-head LEACH)

In LEACH, due to random selection of cluster-heads it is quite possible that selected cluster-head has not had enough
energy to transmit data to base station and dies. So, all the data that lies within cluster-head lost. V-LEACH provides solution
to this problem by making an extra cluster-head called vice cluster-head [20–22]. In V- LEACH:

• Cluster-head performs same function as in LEACH i.e. collect, compress and transmit data to base station.
• If cluster-head dies vice cluster-head perform duty of cluster-head.

Cluster-head and vice-cluster-head selection are done on basis of energy, distance and residual energy. But this protocol
also does not provide any solution for problem where vice cluster-head dies, results in increase of energy dissipation and
hence decreasing network life time. Topology of V-LEACH is shown in Fig. 2 as:
Due to distribution of load among cluster-head and vice cluster-head, V-LEACH has 49% more network life time as
compared to LEACH protocol for a designed network of size 100 m × 100 m with 100 nodes; with each node have an initial
energy of 0.5J. The designed network had a cluster-head probability of 20%, base station at distance of 50 m away from filed,
ETX 50 pJ, ERX 50 pJ, EDA 5 pJ.

4. Other heirarchial routing protocols

4.1. PEGASIS (Power efficient gathering sensor information system)

PEGASIS protocols aims at increasing network life-time of network by increasing energy efficiency and uniform energy
consumption across network nodes. Further, it reduces the delay that occurs to data during their way to base station. PEGASIS
assume that every node has global knowledge about other nodes positions. As compared to other protocols, PEGASIS uses
chain structure for data gathering and dissemination [23–25]. In this protocol, every node work as receiver as well as
transmitter to close neighbours. This means every node sends and receives one packet in each round and acts as leader of
the chain one time among the other nodes of the network.
Fig. 3 shows construction of chain in a network. In this node 0 connects to closes neighbour node 1, node 1 find closest
neighbour node 2 and neighbour node 2 connects to node 3. If a node among these dies, then the chain is reconstructed
without the dead node. Every node except last node in chain fuses own data with its neighbour’s data and then pass that to
next neighbour in the chain. In the above figure node 0 will send its data to node 1. Node 1 fuses own data with this received
data and forwarded this data to node 2 and pass on this data to node 3. Finally, node 3 transmits whole data to BS. PEGASIS
ensures that a relatively far away node did not become leader of chain, which make node to consume more energy for
transmission in every round and hence decreasing network life time. This condition was implemented by setting a threshold
value on neighbour distance for any node to become leader of chain. Further, an improvement was achieved on PEGASIS
that evaluates the threshold neighbour-distance along with the threshold energy-level of nodes those offer themselves to
act as a leader. PEGASIS performs better than LEACH due to elimination of overhead of dynamic cluster formation, reduction
of transmission distance between non cluster-head nodes and transmission of data from chain leader to base station in one
6596 V.K. Arora (Research Scholar) et al. / Optik 127 (2016) 6590–6600

Fig. 4. Hierarchical clustering in TEEN and APTEEN Protocols [26,27].

go only. In a network of 100 randomly distribute nodes, PEGASIS protocol prolonged the life-span of the network two times
as compared to LEACH. The results are reported for a network size of 50 m × 50 m with base station located at distance of
25,150 m. In second scenario, it has three times more life as compared to LEACH when 1%, 20%, 50% and 100% of nodes died
in a network of size 100 m × 100 m with base station located 50,300 m away from the field area. This behaviour is due to the
uniform distribution of load among all the nodes in the network (Fig. 4).

4.2. TEEN (Threshold sensitive energy efficient sensor network)

TEEN is a first reactive cluster based hierarchical routing protocol in which the cluster-head is also responsible for
broadcasting hard threshold and soft threshold values to its member nodes [26]. In this scheme, the nodes keep on sensing
the environment continuously and start transmitting sensed data to the cluster-head only when attribute cross its hard
threshold value. Initially, if sensed attribute reaches its hard threshold; this value is stored in the parameter called as sensed-
value (SV) parameter. Further, node will transmit data if it finds sensed value is either greater than the hard threshold-value
or the difference between the sensed-value and SV parameter is more than or equal to soft threshold. This leads to decrease
in number of data transmission attempts and saves energy. In addition to it, soft threshold also decreases the numbers of
transmissions. It did not allow any further transmissions if the sensed value is smaller than the broadcasted soft threshold
value. Major limitation of this protocol is if sensed attribute never reaches its threshold, node will not send any data to
cluster-head i.e. there will be no communication and hence, the user will be unaware of the death of all nodes. So, TEEN is
never used where application requires data constantly. Furthermore, it has been observed that TEEN (soft-threshold) offered
energy dissipation at half rate as compared to LEACH for a designed network of 100 nodes having initial energy of 2 J placed
randomly in filed size of 100 m × 100 m due to periodic transmission to base station.

4.3. APTEEN (Adaptive threshold TEEN)

Both TEEN and APTEEN uses two level clusters to send their data to base station as shown in Fig. 3. Basically APTEEN is an
enhancement over TEEN protocol and designed especially for hybrid networks in which cluster-head broadcasts following
parameters to their member nodes [27]:

• Physical parameters
• Threshold values.
• TDMA schedule.
• Maximum time between consecutive reports.

In TEEN, TDMA scheduling ensures that no collision occurs during data transmission from member nodes to cluster-
head, but causes delay. APTEEN resolves this problem by modifying TDMA schedule by adding CDMA schedule to it [27].
This modification allows sleeping nodes to send their data followed by idle nodes. APTEEN helps in historical analysis of
previous data values, monitoring of an event and snapshot of the current network. The author has evaluated the performance
of APTEEN with LEACH and TEEN in terms of energy consumption and no of alive nodes [27]. And found that performance of
APTEEN lies in between LEACH and TEEN because APTEEN transmits data periodically whereas TEEN transmits time critical
data.

5. Improvement over heirarchail routing protocol

5.1. DFCA (Distributed fault tolerance clustering algorithm)

In DFCA, gateways have more energy than the normal node and made as cluster-heads [28]. These special nodes are also
battery operated and hence, limited life time. Consequently, their proper usage in network increases the life-time of the
network. DFCA also implemented fault tolerance to tackle the death of gateways [28]. A sensor node select nearby gateways
on the basis of: remaining energy of gateway, distance of node to gateway and distance of gateway to base station. Nodes are
 V.K. Arora (Research Scholar) et al. / Optik 127 (2016) 6590–6600 6597

Table 6
Comparison of DFCA between 100 v/s 360 nodes over different parameters.

Parameters DFCA

Area 300 × 300 100 × 100

No of nodes 360 100
Covered Nodes 290 69
No of gateways 36 9
Nodes having backup 34 22
Energy Left 120.1256 28.9414
Routing Overhead 9.0580 5.8500
Number of Rounds When First Node Dies 1398 1238

Table 7
Comparison of EDFCA between 100 v/s 360 nodes over different parameters.

Parameters EDFCA

Area 300 × 300 100 × 100

No of nodes 360 100
Covered Nodes 310 84
No of gateways 36 7
Nodes having backup 14 22
Energy Left 160.2568 43.1022
Routing Overhead 2.8540 0.8400
Number of Rounds before First Node Dies 1470 1267

distributed randomly in WSN and due to random deployment cluster-heads may have unequal number of nodes and some
nodes may be left uncovered. These uncovered nodes communicate with the cluster-head indirectly (by using cluster-head
of the covered set). But as number of uncovered nodes increases, energy to find cluster-head increases and hence network
life-time decreases.
In DFCA, base station assigns unique identity to all nodes including gateways in the network and then sends HELLO
messages (consisting of gateway identity, residual energy and distance to base station) to each gateway. From the strength
of received HELLO message gateway compute their distance from base station. As in LEACH, sensor nodes join nearest
gateway and every gateway made their own clusters. After completion of cluster formation, steady phase starts in which
gateways starts receiving data from sensor nodes of their cluster, aggregate them and send them to base station using TDMA
scheduling.
Fault tolerance is implemented during steady state. If sensor nodes of any cluster did not receive any data or any message
from gateway that means some fault occurs at gateway. So, all nodes of this cluster become inactive nodes. This can be
confirmed by broadcasting HELP message to nodes of neighboring nodes. If reply comes from the sensor node of the neighbour
cluster then that inactive nodes start sending their data to neighbour gateway through neighboring node(s) called as backup
node (s). Later, inactive node join’s this new gateway and start sending data to new gateway directly i.e. without using
backup node. But the load over the newly joined gateway tends to increases which ultimately decreases life of that gateway.
The author has investigated two networks over size of 100 m × 100 m and 300 m × 300 m consisting of 100 and 360 nodes
with 9 and 36 gateways respectively [28]. In both networks each node has an initial energy of 2 J whereas each gateway has
an initial energy of 10 J. Table 6 shows the performance of DFCA algorithm over these two networks.
From Table 6, it is clear that DFCA performs better with 300 nodes randomly distributed over area of 300 m × 300 m than
network of 100 nodes randomly distributed over field size of 100 m × 100 m although both are fed with same initial energy.
This means in spite of catering more number of nodes in larger area, there are lesser overhead spent on searching backup
nodes using HELLO messages (again an overhead or source of energy wastage). Also, more number of back up nodes in this
network means easier availability of backup nodes, results in lesser energy consumption for searching of backup node.
Further, an improvement over DFCA named as EDFCA is reported in which whole area is divided into grids. In these
grid-areas, the nodes are randomly distributed but with the condition that every grid should have equal number of nodes.
Main advantage of implementing grid approach is that it increases the covered nodes, decreases broadcast messages sent
by uncovered nodes for finding new gateway. So EDFCA decreases the routing overhead and hence decrease the energy
consumption for routing.
EDFCA also improved fault tolerance algorithm of DFCA (Distributed Fault Tolerance Clustering Algorithm). In DFCA,
whenever the gateway of any cluster is dead, then the inactive nodes of that cluster joins neighboring gateway and start
sending their data to base-station using this new gateway. But the load over the newly joined gateway increases and by
virtue of this network life-time decreases.
This limitation is removed in EDFCA by making maximum residual energy node among all inactive nodes as new cluster-
head and forwarding member nodes data to base station using this new cluster-head.
EDFCA evaluated two networks of size 100 m × 100 m (divided in grid of 3 × 3) and 300 m × 300 m (divided in grid of
6 × 6) consisting of 100 and 360 nodes with one gateway for every grid i.e. 9 and 36 respectively. In both networks each
6598 V.K. Arora (Research Scholar) et al. / Optik 127 (2016) 6590–6600

node was fed with initial energy of 2 J and each gateway was fed with initial energy of 10 J. Table 7 shows the performance
of EDFCA over these two networks:
It is also clear from Table 7 that EDFCA shows better results for larger network with number of nodes. Also, it outperforms
DFCA because of lesser uncovered nodes and more number of backup nodes. This results in lesser energy consumption, would
incurred on searching covered node and backup node.

5.2. TSEP (Threshold sensitive stable election protocol)

TSEP is a reactive protocol in which nodes continuously sense environment but transmit only when threshold of different
parameters (set by user) is reached [29]. Main disadvantage of this protocol is sensed parameters did not cross threshold
values, used did not get any information. It has three different types of nodes: normal nodes, intermediate nodes and advance
nodes. Advance nodes have maximum energy, nodes having minimum energy are normal nodes and nodes having energy
between advance nodes and normal nodes are called as intermediate nodes. TSEP selects cluster-head on principle of LEACH
i.e. if generated value is less than threshold then this node become CH. But threshold is calculated on following basis:
⎧ Pnrm
⎨   ifNnrm ∈ G
1
Tnrm = 1 − Pnrm r.mod
⎩ Pnrm
0otherwise

⎧
⎪
⎨ Pint
  ifNint ∈ G
1
Tint = 1 − Pint r.mod
⎪
⎩ Pint
0otherwise

⎧
⎪
⎨ Padv
  ifNadv ∈ G
1
Tadv = 1 − Padv r.mod
⎪
⎩ Padv
0otherwise

where G , G and G are normal, intermediate and advance nodes that have not become CH in the previous rounds. After
cluster-head selection, cluster-head broadcast two values: hard threshold and soft threshold. Nodes will not transmit data
packets until and unless sensed value does not reach hard threshold. As we know that soft threshold is minimum change
that node should observe, so it further reduces number of transmission, results in energy saving.
RFLSEP is an improvement of TSEP, in which level of heterogeneity is increased from three to five namely—normal nodes,
advance nodes, intermediate nodes, super nodes and ultra nodes. Hierarchy of nodes with respect to energy is ultra nodes,
super nodes, advance nodes, intermediate nodes and normal nodes. All types of nodes are 0.2% of total nodes except for
normal nodes. In this protocols, the threshold for all five different types of nodes are being deliberated as
⎧ Pnrm
⎨   ifNnrm ∈ G
1
Tnrm = 1 − Pnrm r.mod
⎩ Pnrm
0otherwise

⎧
⎪
⎨ Pint
  ifNint ∈ G
1
Tint = 1 − Pint r.mod
⎪
⎩ Pint
0otherwise

⎧
⎪
⎨ Padv
  ifNadv ∈ G
1
Tadv = 1 − Padv r.mod
⎪
⎩ Padv
0otherwise

⎧
⎪
⎨
Psup
  ifNsup ∈ G
1
Tsup = 1 − Psup r.mod
⎪
⎩ Psup
0otherwise
 V.K. Arora (Research Scholar) et al. / Optik 127 (2016) 6590–6600 6599

Table 8
Comparison of first node and last node dead in TSEP and RFLSEP.

Area(m2 ) No of Nodes TSEP RFLSEP

First node dies Last node dies First node dies Last node dies

100 × 100 100 2215 8831 2487 15566

100 × 100 250 2346 8934 2486 15465
250 × 250 250 1593 8798 2405 15424
250 × 250 100 1165 9139 2246 15768

⎧
⎪
⎨ Pult
  ifNult ∈ G
1
Tult = 1 − Pult r.mod
⎪
⎩ Pult
0otherwise

where, G , G , G , G and G””’ are normal, intermediate, advance, super and ultra nodes that have not become CH in the
previous rounds. The proposed work investigated two protocols TSEP and RFLSEP over varying number of network size and
number of nodes, where each node was fed with initial energy of 0.5 J. Table 8 shows the comparison of first node and last
dead (rounds) in these two protocols and found that RFLSEP always outperforms TSEP. We also found that in both protocols,
as we increase the field area for same number of nodes, death of first node occurs early because nodes have to spent large
energy to transmit their data to base station. Also, if we increases number of nodes over same field area then life of nodes
increases because they have to cover lesser distance (due to denser network) and that helps in energy saving.
In the last, a number of other optimization based algorithms are also investigated to enhance the network life-span for
instance by combining the intra-node spatial distance with the rate of battery drain out with respect to signal transmission
using modified ant colony optimization algorithm (mACO) approach [30–32].

6. Conclusion

This work presented different types of hierarchical routing protocols for wireless sensor networks. In the beginning, LEACH
protocol was defined to increase the network life-time by rotating the role of data collection, aggregation and transmission
to offer uniform distribution of load among all nodes. But, LEACH positioned the cluster-head randomly inside the network
that restricted the prolonged network life-span. Further, to overcome this issue, the work is extended to diverse variants
of LEACH for instance C-LEACH, MOD-LEACH, SEP, MH-LEACH, V-LEACH and TL-LEACH. It is observed that SEP outperforms
as compared to LEACH, C-LEACH, and MOD-LEACH as the advance nodes consistently contribute to act as cluster-head that
leads to less energy dissipation. Furthermore, it is computed that if nodes are fed with initial energy of 1 J (instead of 0.25 J and
0.5 J), MOD- LEACH outperforms among the other protocols. This is due to the overhead replacement of current cluster-head
with new one is more that enhances the energy consumption. On introduction of PEGASIS, it has been observed that it offers
prolonged network life-span about two times more than LEACH in the network. Further, it is revealed out the on increasing
the network area in conjunction with the distance of base-station from field area, network life-span improves three times
the LEACH. The work is extended to investigate TEEN and APTEEN hierarchical routing protocols and observed that after
incorporating CDMA scheduling rather than TDMA scheduling, APTEEN performs in between LEACH and TEEN in terms of
energy consumption. In the last, we proposed grid based EDFCA algorithm by dividing network into grid consisting of equal
number of nodes. Each node having high initial energy relatively in a grid acts as a cluster-head and named as Gateway. The
proposed scheme is the modification pre-existed DFCA protocol. The obtained observations revealed out that the proposed
EDFCA exhibits high network life-time, lesser routing overhead, lesser energy consumption.

References

[1] C.S. Raghavendra, K.M. Sivalingam, T.Z. Eds, Wireless Sensor Networks, Kluwer Academic, New York, 2004.
[2] E. Cayirci, R. Govindam, T. Znati, M. Srivastava, Wireless sensor networks, Int. J. Comput. Telecommun. Netw. 43 (4) (2003).
[3] T. Znati, C. Raghavendra, K. Sivalingam, Special issue on wireless sensor networks, guest editorial, Mob. Netw. Appl. 8 (August (4)) (2003).
[4] J. Gnanambigai, N. Rengarajan, K. Anbukkarasi, Leach and its descendant protocols: a survey, Int. J. Commun. Comput. Technol. 1–3 (2) (2012).
[5] N.A. Pantazis, S.A. Nikolidakis, D.D. Vergados, Energy-Efficient Routing Protocols in WSN: A Survey, IEEE Communications Surveys & Tutorials, vol 15,
No 2, second quarter 2013,1553–877 X/13 (C) 2013 IEEE, pg no 551–588, 2013.
[6] J. Arce, A. Pajares, O. Lazaro, Performance evaluation of video streaming over Ad hoc networks of sensors using FLAT and hierarchical routing
protocols, book, Mob. Netw. Appl. (2008) 324–336.
[7] W.R. Heinzelman, A. Chandrakasan, H. Balakrishnan, Energy-efficient communication protocol for wireless microsensor networks, Proceedings of
the 33rd Hawaii International Conference on System Sciences (2000) 1–10.
[8] M.J. Handy, M. Haase, D. Timmermann, Low Energy Adaptive Clustering Hierarchy with Deterministic Cluster-Head Selection, Proceedings of 4th
International Workshop on Mobile and Wireless Communications Network, USA,Vol. 1, pp. 368–372, 2002.
[9] N. Bulusu, J. Heidemann, D. Estrin, GPS-less Low Cost Outdoor Localization for Very Small Devices, Technical Report 00-729, Computer science
department, University of Southern California, 2000, Apr.
[10] A. Savvides, C.-C. Han, M. aind Srivastava, Dynamic fine-grained localization in Ad-Hoc networks of sensors, in: Proceedings of the Seventh ACM
Annual International Conference on Mobile Computing and Networking (MobiCom), July, 2001, pp. 166–179.
6600 V.K. Arora (Research Scholar) et al. / Optik 127 (2016) 6590–6600

[11] R. Kaur, D. Sharma, N. Kaur, Comparative analysis of leach and its descendant protocols in wireless sensor network, Int. J. P2P Netw. Trends Technol.
3 (1) (2013).
[12] W.R. Heinzelman, A. Chandrakasan, H. Balakrishnan, An application specific protocol architecture for wireless microsensor networks, IEEE Trans.
Wirel. Commun. 1 (October (4)) (2002) 660–670.
[13] D. Mahmood, N. Javaid, S. Mahmood, S. Qureshi, A.M. Memon, T. Zaman, MODLEACH: a variant of LEACH for WSNs, eighth international conference
on broadband, wireless computing, Commun. Appl. (2013) 158–163, http://dx.doi.org/10.1109/BWCCA.2013.34.
[14] G. Smaragdakis, I. Matta, A. Bestavros, SEP: A Stable Election Protocol for clustered heterogeneous wireless sensor networks, pp. 1–11.
[15] A.S. Neto, A.R. Cardoso, J. Celestino, MH-LEACH: a distributed algorithm for multi-hop communication in wireless sensor networks, ICN, The
Thirteenth International Conference on Networks (2014) 55–61, ISBN: 978-1-61208-318-6.
[16] S. Taruna, Rekha Kumawat, G.N. Purohit, Multi-hop clustering protocol using gateway nodes in wireless sensor network, Int. J. Wirel. Mob. Netw. 4
(4) (2012) 169–180.
[17] M. Usha, N. Sankarram, A survey on energy efficient hierarchical (Leach) clustering algorithms in wireless sensor network, Int. J. Innov. Res. Comput.
Commun. Eng. 2 (Special Issue 1 March) (2014).
[18] Z. Fui, W. Wei, A. Wei, An energy balanced algorithm of LEACH protocol in WSN, Int. J. Comput. Sci. 10 (1) (2013) 354–359.
[19] H. Peng, H. Dong, H. Li, LEACH protocol based two-level clustering algorithm, Int. J. Hybrid Inform. Technol. 8 (10) (2015) 15–26.
[20] N. Sindhwani1, R. Vaid, V LEACH: An Energy Efficient Communication Protocol for WSN, Mechanica Confab ISSN: 2320–2491, vol 2, no 2,
February–March 2013.
[21] A. Ahlawat, V. Malik, An Extended Vic-Cluster Selection Approach to Improve V-LEACH Protocol In WSN, 2012 Third International Conference on
Advanced Computing & Communication Technologies, 978-0-7695-4941-5/12 $26.00 © 2012 IEEE DOI 10.1109/ACCT.2013.60, pg no 236–240, 2012.
[22] H. Shah, S.R. Bhoyar, Improved V-Leach Protocol in Wireless Sensor Network with Data Security, OSR Journal of Electronics and Communication
Engineering (IOSR-JECE) e-ISSN, 2278–2834, p- ISSN: 2278–8735. Volume 9, Issue 5, Ver. II (Sep ? Oct. 2014), PP 49–54 www.iosrjournals.
[23] D. Dekivadiya, Power aware routing protocols in WSNs, Department of computer science and engineering, April 2012. www.iosrjournals.
[24] S. Lindsey, C.S. Raghavendra, PEGASIS: Power-Efficient GAthering in Sensor Information Systems, pp. 1–7.
[25] P. Manimala, R.S. Selvi, A survey on leach-energy based routing protocol, Int. J. Emerg. Technol. Adv. Eng. 3 (December (12)) (2013) 657–660.
[26] A. Manjeshwar, D.P. Agrawal, TEEN: A Routing Protocol for Enhanced Efficiency in Wireless Sensor Networks, 0-7695-0990-8/01(C) IEEE 2001.
[27] A. Manjeshwar, D.P. Agrawal, APTEEN: A Hybrid Protocol for Efficient Routing and Comprehensive Information Retrieval in Wireless Sensor
Networks, Proceedings of the International Parallel and Distributed Processing Symposium (IPDPSí02) 1530–2075/02 $17.00(C) 2002 IEEE.
[28] M. Azharuddin, P. Kuila, P.K. Jana, A distributed fault-tolerant clustering algorithm for wireless sensor networks, International Conference on
Advances in Computing, Communications and Informatics (ICACCI) (2013) 997–1002.
[29] A. Kashaf, N., Javaid, Z. A. Khan, I.A. Khan, TSEP: Threshold-sensitive Stable Election Protocol for WSNs.
[30] V. Sharma, A. Grover, A modified ant colony ptimization algorithm (mACO) for energy efficient wireless sensor networks, Opt.—Int. J. Light Electron
Opt. 127 (February (4)) (2016) 2169–2172.
[31] C. Camilo, J.S. Carreto, An Energy-Efficient Ant-Based Rout-ing Algorithm for Wireless Sensor Networks, 4150, Springer Lect. Notes Comput. Sci.,
2006, pp. 49–59.
[32] J. Kim, S. Lee, B. Cho, Discrimination of battery characteristics using dis-charging/charging voltage pattern recognition, in: Proceedings of the IEEE
Conference on Energy Conversion Congress and Exposition, San Jose, CA, (Sep.), 2009, pp. 1799–1805.