Content uploaded by Nadeem Javaid
Author content
All content in this area was uploaded by Nadeem Javaid on Dec 12, 2019
Content may be subject to copyright.
Received: 26 January 2019 Revised: 2 September 2019 Accepted: 13 October 2019
DOI: 10.1002/ett.3805
SPECIAL ISSUE ARTICLE
NADEEM: Neighbor node approaching distinct
energy-efficient mates for reliable data delivery in
underwater WSNs
Nadeem Javaid
Department of Computer Science,
COMSATS University Islamabad,
Islamabad 44000, Pakistan
Correspondence
Nadeem Javaid, Department of Computer
Science, COMSATS University Islamabad,
Islamabad 44000, Pakistan.
Email: nadeemjavaidqau@gmail.com
Abstract
In this research work, we propose three schemes: neighbor node approach-
ing distinct energy-efficient mates (NADEEM), fallback approach NADEEM
(FA-NADEEM), and transmission adjustment NADEEM (TA-NADEEM). In
NADEEM, immutable forwarder node selection is avoided with the help of
three distinct selection parameters. Void hole is avoided using a fallback recov-
ery mechanism to deliver data successfully at the destination. The transmission
range is dynamically adjusted to resume greedy forwarding among the network
nodes. The neighbor node is only eligible to become forwarder when it is not
a void node. Additionally, linear programming–based feasible regions are com-
puted for an optimal energy dissipation and network throughput improvement.
Extensive simulations are conducted for three parameters: energy, packet deliv-
ery ratio (PDR), and fraction of void nodes. Further, an analysis is performed by
varying transmission range and data rate for energy consumption and fraction
of void node. The results clearly depict that our proposed schemes outperform
the baseline scheme (GEDAR) in terms of energy consumption and fraction of
void nodes.
1INTRODUCTION
In the last decade, UWSNs have gained a lot of attention fromresearchers because of their potential to monitor underwater
environment. UWSNs have an extensive variety of applications such as military defense, monitoring aquatic environment,
disaster prevention, oil/gas extraction, offshore exploration, commercial and scientific purposes, etc.1,2
In underwater atmosphere, protocols of terrestrial wireless sensor networks (TWSNs) cannot work perfectly. The TWSN
and UWSN have differences in many aspects such as use of acoustic links instead of radio links. UWSN topology is more
dynamic than that of TWSNs because nodes move freely with water currents and change their position frequently, local-
ization of nodes is difficult as compared with TWSNs, deployment of nodes is comparatively sparse, energy is limited, and
after deployment, it is difficult to recharge the batteries. In addition, UWSNs face many challenges like low bandwidth,
high propagation delay, and high communication cost.3
Various routing protocols are proposed to minimize the energy consumption and to avoid void occurrence.4-6 The void
node is defined as the node that fails to deliver its information towards the destination because of the unavailability of
the forwarder node. To recover data from the void node, a fallback recovery mechanism and transmission adjustment
approaches are widely used. In the fallback method, the void node looks for every alternate route in its communication
range to resume the greedy forwarding. While transmission adjustment is the process of changing the transmission power
dynamically to bypass the void communication area and to find an immediate forwarder node for carrying out the data
Trans Emerging Tel Tech. 2019;e3805. wileyonlinelibrary.com/journal/ett © 2019 John Wiley & Sons, Ltd. 1of22
https://doi.org/10.1002/ett.3805
2of22 JAVAID
communication among the network nodes. However, these two approaches follow greedy forwarding where every node
finds the node that is closest to the sink to save energy. This greedy approach selects node based on a single parameter
such as distance and cause an immature death of the intermediate node. Thus, to avoid the aforementioned immutable
forwarder selection, multiple parameters need to be considered for cyclic selection of the node to avoid the sudden death
of the forwarder node. Therefore, an efficient protocol is required, which can minimize energy consumption and be able
to avoid void node occurrence.
To minimize energy consumption, geographic routing is used because of its simplicity and scalability. Moreover, it does
not require to maintain complete route from a source to a destination.3Whereas, opportunistic routing is exploited to
transmit data reliably by selecting multiple neighbor nodes at each hop. This set consists of nodes that are aligned with
respect to certain priorities. Only the node with the highest priority is able to forward the data packet. Other nodes will
cancel their transmissions when they are acknowledged that the data packet is delivered successfully.
However, the geographic routing paradigm causes void hole problems because of the selection of only one node for
data transmission. On the other hand, the opportunistic routing results in redundant transmission and consumes more
energy.7Thus, to achieve multiple objectives at the same time, both schemes are combined to improve energy efficiency
and avoid void node occurrence in the network.
1.1 Motivation and contribution
Motivated by above consideration, we propose neighbor node approaching distinct energy-efficient mates (NADEEM)
and its two variants with special features of void recovery using fallback and transmission adjustment approaches. The
NADEEM uses geographic and opportunistic routing for energy-efficient data communication. While in transmission
adjustment NADEEM (TA-NADEEM), the transmission of the void node is adjusted to bypass the void region and
forwards the data to the immediate available node. While fallback recovery NADEEM (FA-NADEEM) uses backward
transmission and finds a different routing path for successful data delivery at the destination. The contributions of our
proposed schemes are given as follows.
The contributions of the proposed work are as follows. First, the NADEEM routing scheme considers three metrics
(energy, depth, and number of neighbors) to select the neighbor node and prioritize the forwarders for data packet trans-
mission at the destination. Second, the FA-NADEEM uses backward transmissions to avoid the void node. It selects a
node that is closest to it in the backward direction and finds route towards the destination. This avoids the void node;
however, the trade-off of energy consumption exists. Third, the TA-NADEEM adjusts the transmission range of the void
node adaptively to forward the data packet at the destination successfully.
The rest of this paper is organized as follows. Section 2 discusses related work about routing protocols in UWSNs and
the problem statement is defined. Section 3 presents the system model and propagation model of the acoustic signal.
In Section 4, the proposed work is presented in detail. In Section 5, the feasible regions of energy minimization and
throughput maximization using linear programming are discussed in detail. We present the performance evaluation of
our schemes using simulation in Section 6. Finally, the performance trade-offs of proposed schemes are presented in
Section 7 followed by our conclusion in Section 8.
2RELATED WORK
In this section, related work on routing protocols in UWSNs is presented based on proposed features, advantages, and
disadvantages. A summary of the contributions of the discussed existing literature is given in Table 1.
Depth-based routing (DBR)8only requires depth information to forward data packets towards the sink with a greedy
approach. Each node forwards the data packet based on depth parameter and suppresses communication to avoid redun-
dant transmission via calculating holding time. The priority is assigned to the forwarder node solely based on depth;
however, it depends that packet is new and not delivered at the destination. The DBR provides improved network life-
time and high data delivery ratio. However, this greedy mechanism with one parameter forces immutable selection of the
forwarder node due to which network nodes get partitioned and resources remain under utilized.
An improved adaptive mobility of courier nodes in threshold-optimized DBR (iAMCTD)9is a location free routing
protocol, which is designed for time-critical applications. It provides an improved network lifetime, minimized end-to-end
delay with the help of mobile courier nodes. However, this scheme results in low throughput, because of avoidance of the
unnecessary transmission.
JAVAID 3of22
TABLE 1 Summary of UWSN routing schemes discussed in related work
Technique Feature (s) Achievement (s) Limitation (s)
DBR8Requires only local depth Improved network lifetime Void holes, increased energy consumption
information and greedy forwarding and packet delivery ratio and high end-to-end delay
iAMCTD9Location free routing protocol specially Improved network lifetime, Void hole problem still exists and communication
designed for time-critical applications minimized end-to-end delay overhead due to control packets exchange
BLOAD10 Data is divided in to fractions, transmission Network lifetime, energy balancing, Energy consumption increases
range is logically adjusted stability period and reduce energy holes
Delay-sensitive schemes1Improved delay-sensitive versions, Minimize end-to-end delay and improve Duplicate packets, high energy consumption
adaptable to time-critical applications performance and network lifetime
ARCR11 Network is divided into clusters and Achieves energy efficiency, maximum Network disconnects when the
mobile nodes are used to collect data network lifetime and load balancing relay nodes are disorganized
from other sensor nodes
E-CARP12 Distributed cross layer reactive protocol Improved network lifetime and Reduced throughput and high
reduced energy consumption path loss due to mobility
HydroCast3Pressure-based routing and efficient Improved packet delivery ratio Low performance and increased
anycast routing algorithms energy consumption
WDFAD-DBR13 Depth-based routing, depth difference of two hops, Reduced energy consumption and Forwarding using two hops never
neighbor node prediction and forwarding area division low probability of void holes eliminates the void hole problem
AHH-VBF14 Location-aware routing protocol and concept of Reduced duplicate packets and unnecessary Void hole problem exists
adaptive virtual pipeline energy consumption is avoided
ORR15 To calculate the best number of forwarders based on Mitigate the negative effects caused This protocol is not low as required, Number
forwarding cost approximation and considers by redundant packet forwarding of forwarders can wake up simultaneously
residual energy while selecting forwarder sets resulting in duplication of packets
ELBAR16 Approach to forward packets in the Enhances the lifetime of network and Longer routing path, end-to-end
occurrence of routing holes reduces the energy consumption. delay increases and more energy depletion
H2-DARP-PM17 Hop-by-hop dynamic addressing–based Improved packet delivery ratio High energy consumption
routing protocol for pipeline monitoring
(ACH)218 Free association mechanism where Minimizing energy consumption and Transmission delay
nodes associate with CHs enhances network lifetime
CBSST19 Cluster-based sleep/wakeup scheduling Reduced energy consumption, enhanced Keeping the same CH throughout the network
technique designed for WSN network lifetime and packet delivery ratio causes problem to network lifetime in future lifetime
UCBNL20 A high efficiency uneven cluster deployment Enhanced packet delivery ratio, less energy Irregular clustering causes a
algorithm Based on network layered for event consumption and improved network lifetime lteration in the network.
coverage in UWSNs
PSO-ECHS21 Energy-efficient CH selection that is based Energy efficiency achieved It works only for homogeneous networks
on particle swarm optimization
EDDEEC22 Enhanced developed distributed energy Shows improved performance with respect to Imbalanced clustering and
efficient clustering stability period, network lifetime and packet reelection increases overhead
delivery ratio
SEEC, CSEEC and CDSEEC23 Sparsity and density–aware algorithms and sink mobility Reduced energy consumption Low packet delivery ratio
4of22 JAVAID
In balanced load distribution (BLOAD),10 the problem of energy holes is tackled using the mechanism of fragmentation
of data packet. The fragments of every data are delivered via direct and multihop transmissions towards the destination.
Moreover, nodes deployed away from the destination are equipped with higher battery, resulting in improved network
lifetime.
Kong et al11 provided an advancement of localization free routing protocols of DBR, ie, energy efficient DBR and
AMCTD as improved delay-sensitive versions. In this paper, the authors made the aforementioned schemes adapt-
able according to the requirements of the applications. These schemes achieve minimum end-to-end delay and less
consumption of energy. However, duplication of packets, transmission loss, and void hole exist.
In an adaptive relay chain routing (ARCR),12 mobile nodes are introduced to retrieve data from energy hole while
keeping the sink fixed at its particular location over the water surface. Moreover, redundant transmissions are avoided
through the formation of clusters in the network, and mobile nodes directly gather a composite packet from the head
node by moving into its vicinity. Thus, it gains energy efficiency by balancing the data load among the network nodes at
the cost of delay.
An energy-efficient channel-aware routing protocol (E-CARP)13 is a distributed cross-layer reactive scheme. It solves
the limitations of conventional CARP by reducing the control overhead among the network nodes during communication.
It improves network lifetime with minimal energy consumption through exclusion of control packets in the network.
However, this scheme achieves less throughput because of high path loss.
In hydraulic-pressure-based anycast (HydroCast)1routing algorithm, the objective of void avoidance is tackled with
consistent broadcasting of control message at the sink. With the availability of recent neighbor information, data packets
are relayed reliably towards lower depth forwarders in the direction of destination. Although the use of gauge apparatus
does provide accurate information because of inhostile acoustic environment. However, it has high throughput at the cost
of more energy consumption.
In weighting depth and forwarding area division–DBR (WDFAD-DBR),14 the depth of current and next forwarding
node in the direction of the sink deployed at the water surface is considered. The information helped in avoiding void
node selection with a less packet drop ratio. However, knowledge up to hops never guarantees reliable forwarder selection
especially when the deployment is sparse.
An adaptive hop-by-hop vector-based forwarding (AHH-VBF)15 uses a pipeline whose radius can be adjusted adaptively
to retrieve data packets from the void region. Additionally, the restriction of forwarding area allows minimum battery
dissipation because of very few redundant transmissions. This scheme improved packet delivery ratio and reduced energy
utilization and delay.
An opportunistic routing based on residual energy (ORR)16 designed for asynchronous duty-cycled wireless sensor
networks. It considers residual energy to compute most energy-efficient forwarder nodes. Moreover, the problems of load
balancing, disconnection in coverage, and immature death of intermediate nodes are considered for optimal network
lifetime. However, asynchronous duty cycling causes additional delay because of switching from active to sleep mode
and vice versa.
In energy-efficient and load-balanced distributed routing (ELBAR),17 Abbas et al enabled the collaboration of sensor
nodes to find out the expected polygon of a particular region in the network. The computed polygon's information is
shared in complete network to avoid data loss. The sensed data from each polygon is collected and forwarded through
the escape route, which surrounds the polygon hole based on computed view angle and hole covering parallelogram.
Moreover, the objective is to transmit data successfully in the presence of void holes via an alternate route. Addition-
ally, this scheme increases the lifetime and minimizes the energy consumption of the network nodes; however, escape
route leads to longer routing path, which increases delay and energy consumption when the sparsity of the network
increases.
A hop-by-hop dynamic addressing–based routing protocol for pipeline monitoring (H2-DARP-PM) was proposed by
Ahmad et al18 for efficient forwarder node selection. It uses dynamic addressing to obtain reliable and most effective
forwarder node in terms of energy consumption. Also, assigns dynamic hop address to each node which contributes in
delivering data at the destination. Although, the PDR is enhanced, however, at the cost of high energy consumption.
An adaptive clustering habit ((ACH)2)is presented by Sasikala and Chandrasekar,19 which proposes the mechanism of
free association of nodes with cluster heads (CHs). The CHs are first elected on the premise of threshold value and then, an
optimal number of CHs is selected based on the distance between each head node. In this way, data load managed among
the head nodes that significantly minimizes the energy consumption. Moreover, it reduces propagation distance and
evades back transmissions to reduce energy dissipation; however, there is high communication overhead in associating
normal nodes with head nodes.
JAVAID 5of22
Cluster-based sleep wake scheduling is performed by assuming initiator nodes.20 The initiator nodes initiate commu-
nication and select the CHs. The CH with high energy is set to active mode, while other nodes are sent to sleep mode.
The transmission is resumed once the head node is nominated, and this decreases energy consumption along with high
lifetime and throughput. However, the immutable selection of CH degrades the network performance.
An irregular formation of clusters is performed using a layered approach for event coverage in the work of Rao et al.21
Moreover, a theoretical analysis obtains the expected value and density of nodes on every layer. After acquiring all the
information, network field is divided into various irregular clusters to distribute data traffic across the network. Addi-
tionally, a recovery strategy is exploited to balance the utilization of the node energy in each cluster. Thus, it achieved
enhanced packet delivery ratio and improved network lifetime.
An energy-efficient cluster head selection using heuristic technique named particle swarm optimization (PSO-ECHS)
is proposed by Javaid et al.22 Two parameters are taken into the consideration for calibration called particle encoding and
fitness function. To improve the performance of the algorithm, various metrics including distance between the clusters,
distance between sinks, and residual energy of nodes are taken into account. Thus, using aforementioned metrics, a
weighting function is derived for all clusters that enabled a high packet delivery ratio with minimal loss of data packets.
An enhanced developed distributed energy-efficient clustering (EDDEEC)23 is proposed to minimize energy consump-
tion. It uses two models: The first one is the heterogeneous model, and the second model is the energy consumption
model. The earlier said models are used to make clusters and elect head nodes based on the computed probability via
fitness function. This algorithm achieved improved performance based on stability period, network lifetime, and packet
delivery ratio.
Three schemes, ie, sparsity-aware energy-efficient clustering (SEEC), circular SEEC (CSEEC), and circular depth–based
SEEC (CDSEEC), are proposed for energy-efficient communication in the work of Heidemann et al.24 In SEEC, two
mobile sinks are deployed in sparse region to collect the information and to reduce the probability of energy hole cre-
ation. In CSEEC, mobile sinks are deployed in a circular network to evade energy hole creation. Likewise, CDSEEC is
proposed in this paper to minimize the consumption of energy. However, these schemes result in low throughput, because
of aggregated data packet transmission through cluster heads.
Problem description. The immutable selection of forwarder nodes based on fixed parameters like depth or energy results
in sudden depletion of node battery. Additionally, it leads to the creation of communication voids in the network.2
Although the void node is recovered via depth adjustment in vertical direction, it dissipates energy even quicker to resume
network operations. Further, after the depth adjustment, reconfiguration of network nodes is required. In case of void
node recovery, the void node is moved to a new depth to resume greedy data forwarding among the network nodes. How-
ever, the movement of nodes to new depth as shown in Figure 1 causes excessive energy consumption and also movement
of void nodes without taking into account the change in topology increases void occurrence in a sparse network. Moreover,
communication overhead increases due to continuous exchange of beacon messages.
To tackle the aforementioned problems, an energy-efficient scheme is required; thus, we propose NADEEM and its two
variants to avoid void hole occurrence and to minimize the energy dissipation with a fall back recovery mechanism and
transmission power adjustment. The details of the proposed work are presented in the following sections.
3PROPOSED SYSTEM MODEL
In our proposed schemes, an acoustic multisink architecture is used to enable communication among sensor nodes.
Nodes are randomly deployed to sense, collect, and transmit data to sink nodes positioned at the surface of water2as
shown in Figure 2. Nodes are of two types: relay nodes and anchored nodes.13 Anchored nodes are static at the bottom of
the ocean surface, whereas relay nodes are deployed randomly in the acoustic environment. The relay and anchor nodes
use acoustic signal to carry information towards the destination. The anchored nodes obtain the sensed information and
forward it towards the sink located at the surface with the help of relay nodes. Whereas, at the surface, sink nodes relay
data through radio signals for transmitting information at the base station. It is assumed that packets received at a single
sink are considered to be received at all sink nodes successfully.
In our work, we made some assumptions because our focus was on the energy efficiency and void avoidance. First, a
consideration has been made that every node (relay, anchored, and sink nodes) can obtain its coordinates with the help of
localization services.7Second, the communication is symmetric, where communication between any two random nodes
results in the consumption of the same amount of energy. For instance, node i transmits data to node j with energy E and
when j sends data to i, the battery dissipation remains the same. It only changes when the distance between two nodes
6of22 JAVAID
FIGURE 1 Illustration of depth adjustment and void hole
problem in GEDAR
FIGURE 2 System model
JAVAID 7of22
vary. Third, all acoustic nodes have the power to adjust communication range autonomously and sink node can receive
multiple packets at the same time without any data loss or collision. Fourth, the effect of water currents is in the direction
of horizontal, while in vertical direction the movement is almost negligible. In the following paragraph an overview of
Thorp's propagation model2has been presented to model the energy consumption according to the distance between the
source and the destination.
The path loss due to unhindered propagation route for a signal having frequency fover a distance lis given as24:
A(l,𝑓)=ls𝛼(𝑓)l,(1)
where sis the spreading factor and 𝛼(f)is the absorption coefficient. The geometry of propagation is described using
the spreading factor kand its values for spherical, cylindrical,l and practical spreading are s=2, s=1, and s=1.5,
respectively. The absorption coefficient 𝛼(f),indB/kmforfin KHz, is described by the Thorp's formula2as
10logA(l,𝑓)=s.10logl+l.10log𝛼(𝑓).(2)
The common signal-to-noise ratio (SNR) over lis given as
SNR(l)= Energyb∕A(l,𝑓)
Noise0
=Energyb
Noise0ls𝛼(𝑓)l,(3)
where Energ𝑦band Noise0are constants that reveal per bit energy transmission and noise power density on a nondeclining
additive white Gaussian noise channel.3As in the work of Noh et al,25 Rayleigh fading is used to model a small-scale
system where SNR has the following distribution probability:
𝜌l(Y)=∫∞
0
1
SNR(l)e−Y
SNR(l).(4)
The probability of error is given as
𝜌e(l)=∫∞
0
𝜌e(Y)𝜌l(Y)lY.(5)
Here, 𝜌e(Y)denotes the random modulation at a specific value of SNR and Yis the probability of error. In this paper, we
use the binary phase shift keying (BPSK) modulation in which each bit is carried by a symbol.26 The error probability
over lis taken from the works of Yang et al27 and Noh et al28:
𝜌e(l)= 1
2
1−SNR(l)
1+SNR(l)
.(6)
The data packet delivery probability for nbits is given by
𝜌(l,n)=(1−𝜌e(l))n.(7)
4PROPOSED WORK
In this section, we describe our proposed schemes in detail. Initially, we will discuss the beacon message dissemination
in the network for configuration. Additionally, neighbor node selection mechanism will be presented for successful data
communication. Moreover, an efficient mate selection is discussed to show its effectiveness for reliable data transmission.
Dissemination of beacon message. The beacon dissemination is vital to enable the configuration of the nodes for suc-
cessful data communication. Once all nodes and sinks are deployed in the network volume, collaboration is important
to relay data in a multihop fashion from a source to a destination. Thus, a beacon is transmitted by a sink node from
the surface of water, which is also equipped with an acoustic modem. This message includes unique information like
identification coordinates (x,y,z). The sinks are equipped with positioning system and have the area facts of each other.2
8of22 JAVAID
Similar is the case for sensor nodes. The beacon transmitted by sensor nodes consists of sequence number, ID, and x,y,
zcoordinates. The positioning system is ineffective in underwater due to excessive frequency signal absorption.2There-
fore acoustic localization services are used for acquiring node information.7Moreover, to reduce the energy consumption
and minimize the communication overhead, the size of the beacon is reduced by only keeping information of node ID
and coordinates. After receiving the information, a node updates its entry in the neighbor table. Beacons are generated
to inform about the location of the sensor node. With this knowledge, neighbor nodes can also access an energy optimal
forwarder node within the communication range. Algorithm 1 represents the process of dissemination of beacon trans-
mission and reception. Each beacon consists of seqnum, its ID, and its dimensions (x,y,z). To avoid protracted sizes of
beacon, source node only considers the x,ydimensions of sink node as shown in lines 5 to 12. At line 20, when the source
node receives the beacon message from the sink node, it updates the entry in that particular sink set Sn; else, it updates
its neighbor node set. Line 25 shows that on every update the source node changes its flag to zero that data is not sent to
the neighbor node. In this way, in each iteration, a beacon message is tracked and according to lines 7 to 10.
4.1 The NADEEM
To ensure reliable data transmission among the network nodes, the forwarder selection is based on multiple parameters:
energy, depth, and number of neighbors. The consideration of three parameters improves the energy consumption and
avoids the void occurrence. The high communication overhead is reduced by only disseminating position information.
First, a set of neighbor nodes is nominated to avoid retransmission of the packet, while most energy efficient mate is
appointed for relaying the packet at the destination, One node transmits a packet at a time to minimize collision and
reduce data packet loss. The details of each phase are given in the following.
Distinct mate selection. An optimal utilization of resources is always desired. In an acoustic environment, a node battery
is one of the most crucial asset that needs to be efficiently consumed during the process of data communication. The void
JAVAID 9of22
occurrence causes network partition; thus, an approach that can avoid the occurrence or bypass it is required. Therefore,
we opt to select various forwarding mates from a neighborhood.29,30 The multiple neighbor selection concept is to ensure
backup when node with the highest priority fails to deliver the desired data packet at the respective destination. The
following mathematical expression is used to find out mates with desired features.
M(i)= E(i)
(Trange ×N)+1.(8)
In the given equation, M(i)shows the mate selected for data communication in the network, while idepicts the number
of mates selected (ivaries from 0 to N,whereNcould be any real positive integer). The right-hand side of the expression
shows a computation mechanism of the priority factor. Here, E(i)denotes the energy of node iand it is multiplied with
Nnumber of neighbors to increase its suitability because higher the neighbor number is, more will be backup available
incase of transmission failure and 1 is added to avoid when Nnumber is 0. The fundamental concept of this approach is
to boost the data packet in the direction of sinks on the surface. The selection of multiple neighbor nodes results in fewer
number of retransmissions of the data packets.27,28
Moreover, one more factor is considered before nominating the forwarder node, ie, packet advancement (ADV),27 to
choose the forwarder node that leads the packet towards the destination. The ADV is defined as the difference between
distance of a source node Siand destination node Diwith difference of distance between neighbor Yand Ni.2njis the
node that has a packet to transmit, its neighbor set is represented as Nj(t)and the set of sinks is represented as Sj(t),where
tis representing the time. Therefore, the neighbor node set in geographic and opportunistic routing is given as
Ni=nk∈N𝑗(t)∶∃sv∈S𝑗(t)Ndn𝑗,s∗
i−Nd(nk,sv)>0,(9)
where S∗
iis from the set of sinks Si(t)and it is the closest sink of node njas
s∗
i=argmin∀s𝑗∈S𝑗(t){Nd(n𝑗,s𝑗)}∕ (10)
Energy-efficient mate selection. The sender node set selection is based on normalized ADV (NADV) as proposed by
Coutinho et al2and Noh et al.3,28 Using NADV, the most appropriate next hop forwarder mate from the neighbor node
set Niis selected. The NADV relates the foremost economical trade-off between the proximity and link cost to choose the
priorities of the eligible nodes. So, the greater the NADV of the node is, the greater will be its priority for selection as the
next forwarder. For every next-hop forwarder node nf∈Ni,NADVis
NADV(n𝑓)=ADV(n𝑓)×𝜌l𝑗
𝑓,n.(11)
Here, ADV(n𝑓)=D(n𝑗,s∗
𝑗)−D(n𝑓,s∗
𝑓)is the nfADV towards the closest sink, s∗
𝑓;l𝑗
𝑓is the distance between source node
njand forwarder node nf,and𝜌(l𝑗
𝑓,n)is the probability of nbits over distance l𝑗
𝑓and it is given in Equation 7.
Then, the node set selection and next hop forwarder set selection assumes fi⊆Niis the set that is formed according
to the priorities of NADV.2,28 As the goal of geographic and opportunistic routing is to find the fi⊆Nito maximize the
expected packet advance (EPA). The EPA of fiis formulated in Equation 11.
EPA(𝑓i)=Σ
k
d=1NADV(nd)Πd−1
𝑗=01−𝜌l𝑗
i,n.(12)
Ultimately, the set with maximum EPA is chosen as the next forwarder node set. The node with the highest precedence
is chosen from the forwarder set as a next hop forwarder. It declares its location records to its neighbor node. If a node is
not decided as a next hop forwarder node for a longer period, its packet is discarded. Algorithm 2 presents the next hop
sender selection. In this algorithm, the first NADV of eligible nodes is calculated corresponding to Equation 10. Then, Ni
is aligned corresponding to the priority of NADV. After that, from line 8 to line 18, it finds the set from candidate set Ni.
Every set in fistarts from the high priority node in Niand gradually includes all the nodes that have transmission range
less than 1
2cr. Every node in set fjis then included in Ci. Each set is expanded while keeping the restriction that every node
do not hear one another. At the end, a set fwith high EPA is chosen as the next hop forwarder.
10 of 22 JAVAID
4.2 The FA-NADEEM
When the sparsity of the network increases, the void node occurrence probability also rises. However, the lack of a recovery
mechanism makes it vulnerable to a high packet drop ratio. Therefore, FA-NADEEM is used to ensure that the packet
from the void region must be recovered and delivered at the destination via an alternate path. Figure 3 illustrates the
mechanism of FA-NADEEM. It is evident that when source node Stransmits the data packet, it obtains the information
up to two hops, along with neighbor information to ensure the suitability of the elected node to proceed with the data
communication. There are four neighbor nodes: n1, n2, n3, and n4. The first hop node is n2 because it is more effective to
relay packet in terms of energy and helps in reducing the hop number. The NADEEM further explores the neighbors of
n2andthenn3 and so on. When the n3 is reached, it encounters a void region. Thus, the fallback approach is adopted in
NADEEM and called FA-NADEEM. Figure 3 shows that n3 declares it void, and instead of dropping the packet, it looks
in its list based on the priority to ensure that it has to traverse the least number of hops in a recovery procedure. Similarly,
n3 falls back and transmits the data to n4 in its communication range. Then from n4, the process of data forwarding
is continued in the direction of sink deployed at the water surface. In the communication period, a fallback recovery
procedure is then adopted to discover different routes, which leads towards the destination. As soon as a node is found,
greedy forwarding is resumed to save node battery. Let us assume a scenario depicted in Figure 3, where node Sforwards
the data packet to the node n2, and when at node n3, it detects the void hole region then it uses the fall back mechanism
to forward the data packet towards node n4.
The primary steps of our proposed scheme FA-NADEEM are presented in Algorithm 3. In the given algorithm, first, if
a node is in a fallback restoration mechanism, new data packets might be queued and the greedy forwarding mechanism
is rearranged to resend these data packets. If the node is not inside the communication void hole region, then it will
forward the sensed information greedily towards the sink node. In any other case, it will transfer to fall back restoration
mechanism for successful data communication among the network nodes.
JAVAID 11 of 22
FIGURE 3 Working
mechanism of FA-NADEEM
4.3 The TA-NADEEM
The TA-NADEEM has the same steps of NADEEM, except the adjustment of transmission to avoid the void hole problem,
which is illustrated in Figure 4. During communication, TA-NADEEM computes the number of neighbors of the for-
warder node after adjusting the transmission range. Let us assume a scenario as depicted in Figure 4, where node Shas to
send a data packet towards the destination, it has no neighbor in its transmission range. Instead of opting for the fallback
procedure, it simply adjusts the transmission range and delivers the data packet directly to the immediate node comes
into its adjusted communication range.
When Sfails to find a neighbor node, it adjusts its transmission range as we mentioned earlier in the system model
that every node has the ability to adjust the transmission range in case of void occurrence to avoid the packet drop. The
immediate forwarder is n1, Sacquires the information about the neighbors of n1 to ensure that the selected node is not
a void. Thus, it looks into the neighbor table of n1 and finds that it has two neighbors at higher depth than n1. Thus,
12 of 22 JAVAID
FIGURE 4 Adjustment of transmission range in TA-NADEEM
n1 is selected as the effective forwarder node to continue with the process of data routing. If there is no neighbor in the
communication vicinity of n1, then further adjustment in the transmission range is done until a suitable forwarder is
located. Additionally, it updates the information in packet queue if there is no report regarding the packet delivery. It
helps in avoiding redundant data packet transmission resulting in less energy consumption, which prolongs the network
lifespan. Moreover, in future, nodes look into the neighbor table and avoids the void occurrence. In this way, extra energy
consumed in transmission adjustment is accommodated.
5FEASIBLE REGION
In this section, linear programming is used for fixing the optimization problem and calculating feasible solution. Lin-
ear programming is the simplest mathematical technique, which is used to achieve optimal results. To achieve optimal
results, an objective function needs to satisfy the defined constraints. In the following sections, energy minimization and
throughput maximization are presented.
5.1 Energy minimization
To minimize energy consumption, the proposition of the linear programming–based mathematical technique is used to
achieve best possible results. This approach begins with an objective function followed by linear constraints. The objective
function must satisfy the constraints in order to achieve best results. The objective function is defined as
minΣrmax
r=1Energytax(r)∀r∈rmax (13)
The linear constraints for energy minimization are given in Equations 12(a)-12(c).
C1∶Etx,Ercv ≤Ei(13a)
C2∶n𝑓≤nmin
𝑓(13b)
C3∶Cr≤Crmax (13c)
Equation (13a) ensures that the energy required for transmission and reception should be less than the initial energy Ei
of the node. Equation (13b) shows the constraint for the selection of the forwarder node, which has the minimum energy
consumption, where nfis the forwarder node with minimum energy consumption. Equation (13c) ensures that to receive
a good quality signal, the data should be transmitted within its transmission range, where Tris the transmission range of
the node and Trmax is the maximum transmission range of the node.
Energytax(r)=Σ
rmax
r=1(Energyconsumed(r)) ∀r∈rmax (14)
JAVAID 13 of 22
01234567
Etx (J)
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
Ercv (J)
P (1, 0.2)
E +E = 0.42
P (4, 0.2)
P (4, 0.1)
P (1, 0.1)
FIGURE 5 Feasible region: Energy minimization
The proposed schemes use Equation (14) to calculate the energy consumption per node, where Energ𝑦consumed includes
both reception and transmission energies of the nodes in the network, ie,
Energyconsumed(r)=Etx +Ercv ∀r∈rmax ,(15)
where
Etx =Ptx Datasize
Datarate .(16)
Etx is the energy consumed during the transmission of data, and Ptx is the transmission power.
Ercv =Prcv Datasize
Datarate (17)
where Ercv is the energy consumed in receiving the data and Prcv is the receiving power.
Graphical analysis. To provide clear visualization of the proposed problem, graphical analysis is presented to compute
all possible values within the feasible region. Assuming Datasize =100B,Datarate =50kb∕s,Ptx ={0.5,1…,2}W,and
Prcv ={0.025,0.05,…,0.1}W, the feasible solution for energy minimization is computed as follows from aforementioned
constraints in Equations (13a)-(13c).
1≤Etx ≤4 (18)
0.1≤Ercv ≤0.2 (19)
1.1≤Etx +Ercv ≤0.42 (20)
The feasible region is plotted in Figure 5 via points extracted from Equations 21a-21c and the points on the boundary of
this feasible region are:
P1(1,0.1)=1.1J
P2(1,0.2)=1.2J
P3(4,0.2)=4.2J
P4(4,0.1)=4.1J
Hence, selecting any value from these points will minimize the energy consumption.
5.2 Throughput maximization
To improve throughput, we formulate the objective function of throughput and take its linear constraints according to
the following expression.
MaximumΣrmax
r=1Thr(r)∀r∈rmax (21)
14 of 22 JAVAID
FIGURE 6 Feasible region: Throughput maximization
0 50 100 150 200 250 300 350 400
BFrw
(KHz)
0
50
100
150
200
250
300
350
400
BNFrw
(KHz)
P(50, 150)
B +B = 350
P (100, 150)
P(50, 250)
Constraints of the objective function are given as follows:
C1∶Etx,Ercv ≤Ei,(21a)
C2∶Etx ≤Ere,(21b)
C3∶Cr≤Crmax,(21c)
C4∶Di𝑗≤Dmax
i𝑗,(21d)
C5∶minΣrmax
r=1Br
Frw.(21e)
Equation (21a) ensures that the energy required for transmission and reception should be less than the initial energy Ei
of the node. Equation (21b) shows the constraint of transmission energy Etx, which must be less than the residual energy
Ere. Equation (21c) ensures that to receive a good quality signal, the data packet ought be transmitted within its maximum
transmission range Crmax,whereTnis the transmission range of the node and Trmax is the maximum transmission range
of the node. Equation (21d) ensures a threshold for distance between sender iand receiver jfor successful communica-
tion. Equation (22) shows restriction that load on forwarder nodes and nodes that have less residual energy ought to be
minimum, where BFrw is the bandwidth of the forwarder nodes.
Graphical analysis. Assume a scenario where a total bandwidth is between 150 KHz to 250 KHz, where BFrw shows
the bandwidth allocated to the forwarding nodes with high residual energy, and BNFrw is the bandwidth assigned to
nonforwarding nodes. The bandwidth Ballocated to BFrw and BNFrw is computed using aforementioned constraints in
Equations 21a to 21d:
50 ≤BFrw ≤100 (22)
150 ≤BNFrw ≤250 (23)
200 ≤BFrw +ENFrw ≤350.(24)
The feasible region is plotted in Figure 6 via points extracted from Equations 22 to 24 and the points on the boundary of
this feasible region are as follows:
P1(50,150)=200KHz
P2(100,150)=250KHz
P3(50,250)=300KHz
Hence, selecting any value from these points will maximize the throughput.
6SIMULATION RESULTS AND DISCUSSION
To evaluate the performance of proposed schemes (NADEEM, FA-NADEEM, and TA-NADEEM), we conducted extensive
simulations against existing scheme GEDAR.3Three parameters: energy consumption per packet, packet delivery ratio,
JAVAID 15 of 22
Parameter Value
Nodes 150-450
Sinks 45
Network volume (m3)1500 ×1500 ×1500
Initial energy (J) 70
Transmission range (m) 250
Transmission power (W) 2
Reception power (W) 0.1
Idle power (W) 0.01
Frequency (kHz) 10
Packet size (bytes) 100
Data rate (kbps) 16, 32, 64, 128
TABLE 2 Simulation parameter of UWSNs
and fraction of void nodes are analyzed in detail. Additionally, our major focus is on the maximization of network lifetime
through efficient energy utilization and avoidance of void occurrence to achieve continuous data communication among
the deployed network nodes. Thus, only analysis of transmission range and data rate is conducted to observe variations
in the results. The definition of each parameter is given as follows:
1. Fraction of void nodes: A node that has the ability to sense, collect, and transmit information, but the unavailability
of any node in its transmission range makes it void node. Thus, the fraction of void nodes is termed as the proportion
of void nodes occurred during node communication.
2. PDR: It is termed as the number of packets transmitted from the network nodes to the amount of data packets
successfully reached the destination (sink).
3. Energy consumption: It is defined as the quantity of a node's battery dissipated in monitoring and exchanging
information without concerning with successful data transmission. It is measured in joules (J).
6.1 Performance control parameters
To perform simulations, the calibration of input parameters, which directly affect the performance of the network system.
The simulations are run for ten times and average is plotted using the line graphs. Moreover, the number of nodes is
randomly deployed in a three-dimensional (3D) network volume of 1500 m×1500 m×1500 mwith an increment of
50 nodes to 150 until it reaches to 450 nodes. Additionally, the transmission range is set to 250 m for comparison, and it
varies in transmission analysis from 150 m to 350 m. Also, the data rate is set to 16 kbps for performance evaluation against
existing scheme GEDAR,3whereas the deviation in energy and void occurrence is observed on 32 kbps, 64 kbps, and
128 kbps. In addition, the payload of the data packet is set to 150 B. The values of energy consumption are: Pt=2W,
Pr=0.1Wand Pi=10mW for transmission, reception, and idle energies, respectively.3The aforementioned parameters
are given in Table 2.
6.2 Fraction of local maximum nodes
Figure 7 illustrates the ratio of void nodes of NADEEM and its variants against the GEDAR. It is evident from the results
that the behavior of all schemes is homogeneous and the void ratio decreases with the increase in node number. Ini-
tially, the highest transmission failure ratio can be observed of FA-NADEEM, which falls significantly up to 250 nodes;
however, again a small rise is evident for 250-300 nodes. Then, again, the amount of void node occurrence decreases to
a certain level, which remains consistent until 450 nodes. The decrease in ratio is because of the fall back mechanism
of FA-NADEEM, which looks for an alternative route from all nodes in the communication range to recover and deliver
data packet successfully at the destination.
NADEEM has fewer void appearances at the end of the node increment because it does not have transmission adjust-
ment and fallback and depth adjustment mechanisms (see Figure 7). This help in saving node battery ultimately leading
to fewer nodes and more number of alive nodes. Thus, when the sparsity of the nodes decreases, the probability of void
occurrence also decreases.
GEDAR and TA-NADEEM show almost identical trend in rise and fall of void proportion during the node commu-
nication. Although, the values are different and, at the end, TA-NADEEM beats the baseline scheme because of direct
transmission adjustment, instead of exchanging beacon message to find out the different route (see Figure 7). When a
different route is not available then the depth of the node is adjusted, these aspects consume a lot of energy and suddenly
16 of 22 JAVAID
FIGURE 7 Comparison of fraction of void nodes
150 200 250 300 350 400 450
Number of nodes
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
Fraction of local maximum nodes
node battery depletes. Whereas, TA-NADEEM adaptively adjusts its transmission power and bypass the void node
successfully.
From 150 to 200 nodes, GEDAR shows sudden decrease in fraction of void nodes (see Figure 7). As GEDAR considers
movement of void node towards its neighbor node without checking whether the neighbor nodes have enough forwarder
nodes or not, it therefore has high fraction of void nodes compared with the proposed schemes.
6.3 Packet delivery ratio
Figure 8 shows that the energy tax gradually decreases with the increase in density of network. The highest PDR is
achieved by FA-NADEEM due to its ability of finding data route from an alternate node within the communication vicin-
ity. When a trap of void node occurs, instead of dropping the packet, the source node proceed with a fallback procedure
and looks in 360◦to resume the data transmission process.
NADEEM depicts the moderate success ratio of packet transmission compared with other schemes in Figure 8. The
performance is neither high nor low, and it operates on static parameters; however, the forwarder selection is not based on
single parameter of depth. The consideration of depth and energy ensures rotation of node constantly for efficient node
battery dissipation.
On the other hand, the performance regarding the success ratio of GEDAR and TA-NADEEM is identical in Figure 8.
The energy consumption is high in both of the methodologies because, in GEDAR, depth adjustment in vertical
direction consumes more energy as compared to that of simple recovery methods. Similarly, the TA-NADEEM opts
towards the adjustment of transmission range; the energy consumption depends on the distance between the source and
FIGURE 8 Comparison of PDR
150 200 250 300 350 400 450
Number of nodes
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Packet delivery ratio
JAVAID 17 of 22
150 200 250 300 350 400 450
Number of nodes
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Energy per data packet per node (J)
FIGURE 9 Comparison of energy consumption
destination. When the transmission power is adjusted, it requires more energy to transmit the signal to overcome the
issues of attenuation and fading in dynamic acoustic environment.
6.4 Energy consumption
The comparison of energy consumption in delivering single packet at the destination of all the schemes is depicted in
Figure 9. The minimum energy consumption can be observed from the graphical representation of TA-NADEEM and
slightly high of GEDAR. The energy consumed for one packet is almost 0.02J of TA-NADEEM, whereas the node battery
dissipation in GEDAR is higher than all the proposed schemes, which is a very small number of nearly 0.12 J.
The energy consumption of FA-NADEEM is comparatively more than the other two proposed schemes and almost the
same with GEDAR (see Figure 9). The reason of more energy is of an alternate route that increases the route length,
resulting in more energy requirement. Although it achieves better PDR (see Figure 8), the utilization of limited resource
battery is still considerably higher.
On the contrary, TA-NADEEM shows minimum energy dissipation in delivering a single data packet. It happens
because the only energy required in adjusting power and factors such as message exchange and depth adjustment are not
involved. Thus, a significant amount of energy is saved in TA-NADEEM. Although the NADEEM has initially low energy
requirements, when node density increases, the collision rate of data packets also maximizes resulting in high packet loss.
6.5 Influence of control parameters
In the following discussion, we analyze the performance of NADEEM and TA-NADEEM schemes based on transmission
ranges, data rate, and payload for energy and void fraction. The input value is changed while all other input variables are
the same.
6.5.1 Influence of various transmission ranges on void occurrence
The relationship of transmission range with transmission due to void node occurrence is shown in Figure 10. The lower
the communication range is, the higher the failure rate will be during the data transmission. When the node density is
150, the void ratio is higher than 0.8, almost 0.7, more than 0.57, and greater than 0.3 and 0.2 in the case of 150, 200, 250,
300, and 350 transmission ranges. The pattern described is homogeneous where the decrease is observed until the node
density reaches 450. At 150 transmission range, the minimum ratio is slightly lower than 0.6 when the 450 nodes are
deployed. Still, the ratio is higher than every other transmission range used for the performance analysis. The minimum
is when the range is 350 and density of nodes is 450. The reason of this increase and decrease trends; when range is
more, the probability of finding neighbor nodes in the transmission is high, and thus voids can be easily avoided. On the
other hand, with a smaller communication region, the success ratio becomes low due to the availability of fewer nodes,
especially in sparse deployment. However, it increases with the increase in the transmission range.
18 of 22 JAVAID
FIGURE 10 Occurrence of void nodes in TA-NADEEM at various
transmission ranges
150 200 250 300 350 400 450
Number of nodes
0
0.2
0.4
0.6
0.8
1
Fraction of local maximum nodes
FIGURE 11 Energy consumption in TA-NADEEM at different
transmission ranges
150 200 250 300 350 400 450
Number of nodes
0
0.05
0.1
0.15
Energy per data packet per node (J)
6.5.2 Influence of various transmission ranges on energy consumption
The comparison of energy dissipation is depicted in Figure 11 of various transmission ranges. As we have discussed, the
phenomenon of void occurs due to transmission ranges. It is concluded that when the communication range is high,
the energy consumption is more because of proportionality of distance and transmission with each other. However, it
is clearly evident that the smaller transmission range consumes less amount of energy at the cost of high void fraction
(see Figure 10).
At node density of 150, the energy consumption is lower than 0.05, and this ratio remains the constant until the node
number 300. However, a very slight increment can be observed until the time the node number approaches 450. On the
contrary, the trend for other transmission ranges is different; a continuous decrease is evident from the results. Although
there is a minor increase in the case of 250 and 300, when energy increased during node number of 250 to 300. After that,
we can observe a decrement pattern for all the schemes.
As we have mentioned earlier, energy is directly proportional to distance; thus, the more communication distance is
between the source and the destination, the higher the energy utilization would be. The fall is seen because when the
number of nodes increases in fixed network volume, the ratio of nodes per meter square also increases, thus leading to
higher success of neighbor availability when helping in quick finding of the forwarder node. This is the reason of decrease
in energy with the increase in the node number.
6.5.3 Influence of data rate on energy consumption
In order to verify the performance of NADEEM, we have conducted simulations on different data rates as illustrated in
Figure 12. It is evident from the results that with a higher data rate, the energy consumption is higher. With high data rates,
the time required to transmit data from the source is very less in case of 128 Kb/s; therefore, the chances of collision and
JAVAID 19 of 22
FIGURE 12 NADEEM:
Energy consumption at
different data rates
retransmission factors are evaded. Thus, the node battery is saved, and the lifetime of the network is optimized. Initially,
the utilization is high compared at 64, 32, and 16 Kb/s because the transmission rate is quicker and the probability of
forwarder node is less. Thus, the drop ratio is high. However, as the node density increases, the consumption drops to a
significant extend (around 0.02 J per packet).
On the other hand, at 128 Kb/s, the energy consumption is very low of approximately 0.01 J as compared to other data
rates (see Figure 12). The major point of focus is that the ratio of energy required to transmit data successfully at the
destination increases while at other rates, the quantity falls. This reason is as the node density increases, the collision rate
increases on the wireless channel, which results in high data packet drop.
6.5.4 Influence of data rate on void occurrence
Figure 13 shows the impact of transmission rate on void occurrence in the network. The relationship between void node
appearance during communication and data rate is analysed at 16, 32, 64, and 128 Kb/s. The minimum ratio is evident
at very low data rate or very high data rate. The effective performance can be observed when the data rate is low because
of dynamic acoustic environment. The inevitable mobility because of water currents plays a vital role in discovering or
vanishing neighbors from the vicinity of the communication range. Moreover, the high case helps in minimizing the
transmission delay and easily data can be delivered at the destination without encountering void regions.
Similarly, a linear decrease is clear that at 64 and 32 Kb/s, the ratio is higher, although the ratio is falling with the
increment of node number. At 64 Kb/s, the ratio is 0.7 and came down to around 4.5. While, 32 Kb/s started with just
over 0.5 and ends at 0.3. Thus, the trend of void ratio reduction is the same from top to bottom, and no irregularities exist;
however, a significant difference can be observed between each transmission rate.
7PERFORMANCE TRADE-OFFS
In this section, we review the performance trade-offs of our proposed schemes: NADEEM, FA-NADEEM, and
TA-NADEEM along with the existing base scheme GEDAR. The GEDAR considers movement of void node without check-
ing whether the neighbor nodes have enough forwarder node or not. The number of retransmissions increases the overall
energy consumption in GEDAR due to high probability of void nodes. However, in dense region, the energy consumption
decreases due to less probability of void nodes. The FA-NADEEM reduces void holes and increases PDR. Also, it selects
nodes with minimum neighbors at the cost of low throughput. The TA-NADEEM reduces the void holes by adaptively
20 of 22 JAVAID
FIGURE 13 NADEEM: occurrence of void nodes at various data
rates
TABLE 3 Performance trade-offs
Scheme Features Achieved parameters Trade-offs
GEDAR 2, 2Geographic and opportunistic routing Void hole avoidance results in High energy consumption
with depth adjustment based topology increased performance and high probability
control for communication Recovery of the network of void nodes
over void regions
NADEEM Geographic and Opportunistic Routing Void hole avoidance High propagation
using Backward Transmission results in increased PDR distance
FA-NADEEM Geographic and Opportunistic Routing Improved network performance Low throughput
using Collision Avoidance
TA-NADEEM Geographic and Opportunistic Routing Improved PDR High energy consumption
using Adaptive Transmission Range in sparse region
adjusting its transmission range; however, it results in increased energy consumption. The performance trade-offs are
shown in Table 3.
8CONCLUSION
In this paper, we have proposed NADEEM, FA-NADEEM, and TA-NADEEM routing protocols to enhance the energy effi-
ciency and minimize the fraction of void node occurrence. The NADEEM minimizes the energy consumption by avoiding
immutable forwarder selection. It also achieves higher packet delivery ratio. Whereas, FA-NADEEM avoids successfully
the void node through sending data from the alternate route. However, the use of alternate path increases the number of
hops, which resulted in more energy consumption. Moreover, it has higher throughput as compared with NADEEM. On
the other hand, when FA-NADEEM fails to find another forwarding path, then the transmission of NADEEM is adjusted
to ensure data delivery at the destination node. Additionally, TA-NADEEM looked for an immediate forwarder node and
its neighbors to ensure that the selected node is not void. It has a low network throughput but a minimum energy con-
sumption for one packet. The results are validated through extensive simulations against fraction of void node, energy
consumed per packet, and packet delivery ratio. The outcome clearly shows that the proposed schemes outperformed
the baseline existing scheme (GEDAR) in terms of energy consumption and void avoidance. Further, to see the impact
of transmission adjustment and data rate, results are obtained which illustrate energy consumption rise when transmis-
sion range is increased and void fraction decreases. On the other hand, when the data rate is varied, it shows an opposite
behavior, ie, with an increase in transmission rate, energy dissipation reduces and the fraction of void increases.
As future work, we plan to improve the delay using heuristic techniques. Additional analysis of payload, packet size,
and number of sinks will be conducted to measure variations in the performance of algorithm. Further, the transmission
range adjustment will be dynamic based on various parameters like link quality, energy, and distance. The aforementioned
factors considered will help in achieving an optimal throughput and network lifetime.
JAVAID 21 of 22
ORCID
Nadeem Javaid https://orcid.org/0000-0003-3777-8249
REFERENCES
1. Javaid N, Jafri MR, Ahmed S, et al. Delay-sensitive routing schemes for underwater acoustic sensor networks. Int J Distrib Sens Netw.
2015;11(3):532676.
2. Coutinho R, Boukerche A, Vieira L, Loureiro AA. Geographic and opportunistic routing for underwater sensor networks. IEEE Trans
Comput. 2016;65(2):548-561.
3. Noh Y, Lee U, Lee S, et al. Hydrocast: pressure routing for underwater sensor networks. IEEE Trans Veh Technol. 2016;65(1):333-347.
4. Kheirabadi MT, Mohamad MM. Greedy routing in underwater acoustic sensor networks: a survey. IntJDistribSensNetw. 2013;9(7):701834.
5. Li N, Martínez J-F, Meneses Chaus J, Eckert M. A survey on underwater acoustic sensor network routing protocols. Sensors. 2016;16(3):414.
6. Han G, Jiang J, Bao N, Wan L, Guizani M. Routing protocols for underwater wireless sensor networks. IEEE Commun Mag.
2015;53(11):72-78.
7. Coutinho RW, Boukerche A, Vieira LF, Loureiro AA. GEDAR: geographic and opportunistic routing protocol with depth adjustment for
mobile underwater sensor networks. Paper presented at: 2014 IEEE International Conference on Communications (ICC); 2014; Sydney,
Australia.
8. Yan H, Shi ZJ, Cui J-H. DBR: depth-based routing for underwater sensor networks. In: NETWORKING 2008 Ad Hoc and Sensor Net-
works, Wireless Networks, Next Generation Internet: 7th International IFIP-TC6Networking Conference Singapore, May 5-9, 2008 Proceedings.
Berlin, Germany: Springer; 2008:72-86.
9. Javaid N, Jafri MR, Khan ZA, Qasim U, Alghamdi TA,Ali M. iAMCTD: improved adaptive mobility of courier nodes in threshold-optimized
DBR protocol for underwater wireless sensor networks. Int J Distrib Sens Netw. 2014;10(11):213012.
10. Azam I, Javaid N, Ahmad A, Abdul W, Almogren A, Alamri A. Balanced load distribution with energy hole avoidance in underwater
WSNs. IEEE Access. 2017;5:15206-15221.
11. Kong L, Ma K, Qiao B, Guo X. Adaptive relay chain routing with load balancing and high energy efficiency. IEEE Sens J . 2016;
16(14):5826-5836.
12. Zhou Z, Yao B, Xing R, Shu L, Bu S. E-CARP: an energy efficient routing protocol for UWSNs in the internet of underwater things.
IEEE Sens J. 2015;16(11):4072-4082.
13. Yu H, Yao N, Wang T, Li G, Gao Z, Tan G. WDFAD-DBR: weighting depth and forwarding area division DBR routing protocol for UASNs.
Ad Hoc Netw. 2016;37:256-282.
14. Yu H, Yao N, Liu J. An adaptive routing protocol in underwater sparse acoustic sensor networks. Ad Hoc Netw. 2015;34:121-143.
15. So J, Byun H. Load-balanced opportunistic routing for duty-cycled wireless sensor networks. IEEE Trans Mob Comput.
2017;16(7):1940-1955.
16. Nguyen K-V, Le Nguyen P, Vu QH, Van Do T. An energy efficient and load balanced distributed routing scheme for wireless sensor
networks with holes. JSystSoftw. 2017;123:92-105.
17. Abbas MZ, Bakar KA, Ayaz M, Mohamed MH, Tariq M. Hop-by-hop dynamic addressing based routing protocol for monitoring of long
range underwater pipeline. KSII Trans Internet Inf Syst. 2017;11(2):731-763.
18. AhmadA,JavaidN,KhanZA,QasimU,AlghamdiTA.(ACH)2: routing scheme to maximize lifetime and throughput of wireless sensor
networks. IEEE Sens J. 2014;14(10):3516-3532.
19. Sasikala V, Chandrasekar C. Cluster based sleep/wakeup scheduling technique for WSN. Int J Comput Appl. 2013;72(8).
20. Yu S, Liu S, Jiang P. A high-efficiency uneven cluster deployment algorithm based on network layered for event coverage in UWSNs.
Sensors. 2016;16(12):2103.
21. Rao PS, Jana PK, Banka H. A particle swarm optimization based energy efficient cluster head selection algorithm for wireless sensor
networks. Wireless Networks. 2017;23(7):2005-2020.
22. Javaid N, Rasheed MB, Imran M, et al. An energy-efficient distributed clustering algorithm for heterogeneous WSNs. EURASIP J Wirel
Commun Netw. 2015;2015(1):151.
23. Sher A, Javaid N, Azam I, et al. Monitoring square and circular fields with sensors using energy-efficient cluster-based routing for
underwater wireless sensor networks. Int J Distrib Sens Netw. 2017;13(7). 1550147717717189.
24. Heidemann J, Stojanovic M, Zorzi M. Underwater sensor networks: applications, advances and challenges. Philos Trans R Soc A Math
Phys Eng Sci. 2012;370(1958):158-175.
25. Noh Y, Lee U, Han S, et al. DOTS: a propagation delay-aware opportunistic MAC protocol for mobile underwater networks. IEEE Trans
Mob Comput. 2014;13(4):766-782.
26. Stojanovic M, Preisig J. Underwater acoustic communication channels: propagation models and statistical characterization.
IEEE Commun Mag. 2009;47(1):84-89.
27. Yang H, Liu B, Ren F, Wen H, Lin C. Optimization of energy efficient transmission in underwater sensor networks. In: Proceedings of the
28th IEEE Conference on Global Telecommunications (GLOBECOM); 2009; Honolulu, HI.
28. Noh Y, Lee U, Wang P, Choi BSC, Gerla M. VAPR: void-aware pressure routing for underwater sensor networks. IEEE Trans Mob Comput.
2013;12(5):895-908.
22 of 22 JAVAID
29. Climent S, Sanchez A, Capella J, Meratnia N, Serrano J. Underwater acoustic wireless sensor networks: advances and future trends in
physical, MAC and routing layers. Sensors. 2014;14:795-833.
30. Yadav S, Kumar V. Optimal clustering in underwater wireless sensor networks: acoustic, EM and FSO communication compliant
technique. IEEE Access. 2017;5:12761-12776.
Howtocitethisarticle: Javaid N. NADEEM: Neighbor node approaching distinct energy-efficient mates for
reliable data delivery in underwater WSNs. Trans Emerging Tel Tech. 2019;e3805. https://doi.org/10.1002/ett.3805
- A preview of this full-text is provided by Wiley.
- Learn more
Preview content only
Content available from Transactions on Emerging Telecommunications Technologies
This content is subject to copyright. Terms and conditions apply.
