Content uploaded by Nadeem Javaid
Author content
All content in this area was uploaded by Nadeem Javaid on Jul 16, 2016
Content may be subject to copyright.
RE-AEDG: Reliable AUV-aided Efficient Data
Gathering Routing Protocol for Underwater Wireless
Sensor Networks
By,
Tayyaba Liaqat
MS-06(EE), 130411008
Supervisor
Dr. Nadeem Javaid,
Associate Professor, Department of Computer Science,
COMSATS Institute of Information Technology, Islamabad.
Co-Supervisor
Dr. Moazam Maqsood,
Assistant Professor, Department of Electrical Engineering.
Institute of Space Technology, Islamabad
Department of Electrical Engineering
Institute of Space Technology, Islamabad
2016
i
RE-AEDG: Reliable AUV-aided Efficient Data
Gathering Routing Protocol for Underwater Wireless
Sensor Networks
A thesis submitted to the
Institute of Space Technology
in partial fulfillment of the requirements
for the degree of Masters of Science in
Electrical Engineering
By,
Tayyaba Liaqat
MS-06(EE), 130411008
Supervisor
Dr. Nadeem Javaid,
Associate Professor, Department of Computer Science,
COMSATS Institute of Information Technology, Islamabad.
Co-Supervisor
Dr. Moazam Maqsood,
Assistant Professor, Department of Electrical Engineering,
Institute of Space Technology, Islamabad
Department of Electrical Engineering
Institute of Space Technology, Islamabad
2016
ii
RE-AEDG: Reliable AUV-aided Efficient Data
Gathering Routing Protocol for Underwater Wireless
Sensor Networks
by
Tayyaba Liaqat
MS-06(EE), 130411008
APPROVAL BY BOARD OF EXAMINARS
_______________
Supervisor’s Name
_____________ _____________
Advisor 1 (Name) Advisor 2 (Name)
iii
Certificate
This is to certify that this thesis “Reliable AUV-aided Efficient Data Gathering Routing
Protocol for Underwater Wireless Sensor Networks” is the original work of the author(s) and has
been carried out under my direct supervision. I have personally gone through all the
data/results/materials reported in the manuscript and certify their correctness/authenticity. I
further certify that the material included in this thesis is not plagiarized and has not been used in
part or full in a manuscript already submitted or in the process of submission in partial/complete
fulfillment of the award of any other degree from any institution. I also certify that the thesis has
been prepared under my supervision according to the prescribed format and I endorse its
evaluation for the award of Masters of Science in Electrical Engineering degree through the
official procedures of the Institute.
__________________
(Dr. Nadeem Javaid)
iv
Copyright © 2015
This document is jointly copyrighted by the author(s) and the Institute of Space Technology
(IST). Both the author(s) and IST can use, publish or reproduce this document in any form.
Under the copyright law, no part of this document can be reproduced by anyone, except
copyright holders, without the permission of the author(s).
v
ABSTRACT
Underwater sensor networks are required to collect the demanded information to fulfill the
humanitarian needs. These networks once deployed remains unattended forever. However, it is
required that the network remains functional for as long as possible. As the time goes on the
deployed network starts portioning due to the random death of nodes. Holes creation is a
troublesome for the underwater wireless networks. Since lots of work has been done to minimize
the energy consumption of network but balanced energy consumption among nodes is also a
vigilance task. Therefore, in this thesis we propose a technique DB-EBH that particularly
enhances the network life time as well as the stability period by observing this aspect of balanced
energy consumption among individual nodes. Our proposed mechanism is a merger of two major
data routing techniques: Hop by Hop and Single Hop transmission. Both of these approaches
alternatively and counterbalances each other's limitations. This hybrid technique attempts to
compensate: the exceptional traffic flow burden over the lower depth nodes due to multihoping
as well as the earlier exhaustion of far off nodes due to direct data hoping towards sink. We
evaluate the performance of our proposed schematic in the contrast to a Hop by Hop scheme.
Simulation results strengthen our claim as our technique appears to balance the network energy
besides enhancing the network stability period.
Besides balanced energy consumption another major issue of underwater wireless sensor
networks (UWSNs) is the data reliability. Whereas, for underwater wireless sensor network
statistics reliability influences energy efficiency. Improved data integrity demands channel
constraints to be courteous enough, which is not a case. Therefore, routing demands the inclusion
of diversity technique in itself. Thus our thesis offers another routing protocol whose key insight
is to oversee the above written issues. Cooperative communication mechanism as a diversity
technique is employed in our proposed protocol. However the drawback of extensive energy
consumption is minimized using the mobility assisted routing. Mobile sinks seems to reduce the
energy dissipation in contrast to static one but this comes at the cost of induced message
latencies. Thus a delay optimization model is proposed in our protocol that attempts to overcome
the limitation triggered by the employment of mobile sink. Besides end to end delay sink
vi
mobility poses another challenge that is protocol overhead. It is caused due to dynamically
varying routes in between the sink and sensor nodes. Therefore, our scheme adopts the concept
of gateway nodes. Gateways are meant to provide the links in between the member nodes and
AUV. Gateways track the recent locality of mobile sink and reduce the control packet overhead.
Efficiency of the mechanism is evaluated through extensive simulations. Furthermore the
achieved results are examined in the reference of two pre existing techniques.
Keywords: Underwater wireless sensor networks, Routing protocol, Energy efficiency,
Network lifetime, Cooperative communication, Diversity, Reliability, Throughput efficiency,
Energy optimization, Delay Optimization.
vii
Contents
Contents viii
1 Introduction 4
1.1 Why underwater sensor networks? . . . . . . . . . . . . . . . . . . . 5
1.2 Types of UWSNs and challenges . . . . . . . . . . . . . . . . . . . . 5
1.3 Applications of UWSNs . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Environmental constraints . . . . . . . . . . . . . . . . . . . . . . . 6
1.5 Communication channel . . . . . . . . . . . . . . . . . . . . . . . . 6
1.6 Relay communication . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.7 Thesisoutline.............................. 8
2 Literature review 9
2.1 Overview of counterpart schemes . . . . . . . . . . . . . . . . . . . 13
3 UWSNs: Communication models 16
3.0.1 Attenuation model . . . . . . . . . . . . . . . . . . . . . . . 17
3.0.2 Transmission loss model: Deep and shallow water . . . . . . 17
3.0.3 Noisemodel........................... 18
4 DB-EBH: Depth-Based Energy-Balanced Hybrid Routing Proto-
col for UWSNs 20
4.1 Summary of chapter . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.2 Motivation................................ 21
4.3 Protocol operation . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.4 DB-EBH: Network model . . . . . . . . . . . . . . . . . . . . . . . 24
4.5 Priority node selection . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.6 DB-EBH: Operational flow . . . . . . . . . . . . . . . . . . . . . . . 26
4.7 Simulation results and discussion: DB-EBH and HBH . . . . . . . . 26
4.7.1 Definition of network performance parameters . . . . . . . . 28
4.7.1.1 Network lifetime . . . . . . . . . . . . . . . . . . . 28
4.7.1.2 Stability time . . . . . . . . . . . . . . . . . . . . . 28
4.7.1.3 Packets to sink . . . . . . . . . . . . . . . . . . . . 28
4.7.1.4 Packets to nodes . . . . . . . . . . . . . . . . . . . 28
4.7.1.5 Residual energy . . . . . . . . . . . . . . . . . . . . 28
viii
4.7.2 Results and discussions . . . . . . . . . . . . . . . . . . . . . 29
4.8 DB-EBH: An inspection . . . . . . . . . . . . . . . . . . . . . . . . 32
4.9 Conclusion of chapter . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5 RE-AEDG: Reliable AUV-aided Efficient Data Gathering Rout-
ing Protocol for Underwater Wireless Sensor Networks 34
5.1 Summary of chapter . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.2 Motivation................................ 35
5.3 Systemmodel.............................. 36
5.4 RE-AEDG: The proposed scheme . . . . . . . . . . . . . . . . . . . 39
5.5 Energy consumption model . . . . . . . . . . . . . . . . . . . . . . . 43
5.6 Possible relays selection . . . . . . . . . . . . . . . . . . . . . . . . 47
5.6.1 Bcformation .......................... 48
5.7 AF cooperative relay model . . . . . . . . . . . . . . . . . . . . . . 49
5.8 Sinktrajectory ............................. 51
5.8.1 Elliptical mobility model . . . . . . . . . . . . . . . . . . . . 53
5.9 RE-AEDG: Optimization models . . . . . . . . . . . . . . . . . . . 54
5.9.1 RE-AEDG: Energy optimization model . . . . . . . . . . . . 54
5.9.1.1 Graphical analysis . . . . . . . . . . . . . . . . . . 56
5.9.2 RE-AEDG: Delay optimization model . . . . . . . . . . . . . 57
5.9.2.1 Graphical analysis . . . . . . . . . . . . . . . . . . 59
5.10 AEDG and ULVRP: Optimization models . . . . . . . . . . . . . . 60
5.10.1 AEDG: Delay and energy optimization model . . . . . . . . 61
5.10.2 ULVRP: Energy optimization model . . . . . . . . . . . . . 61
5.10.3 ULVRP: Delay optimization model . . . . . . . . . . . . . . 62
5.11 LVRP and U-LVRP:
A comparative analysis . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.12 Simulation results and discussions: RE-AEDG, AEDG, U-LVRP . . 66
5.12.1 Evaluation metrics . . . . . . . . . . . . . . . . . . . . . . . 66
5.12.1.1 Network lifetime . . . . . . . . . . . . . . . . . . . 66
5.12.1.2 Stability period . . . . . . . . . . . . . . . . . . . . 67
5.12.1.3 PAR.......................... 67
5.12.1.4 Throughput rate . . . . . . . . . . . . . . . . . . . 67
5.12.1.5 E2E delay . . . . . . . . . . . . . . . . . . . . . . . 68
5.12.2 Results and discussions . . . . . . . . . . . . . . . . . . . . . 68
5.12.3 Relative inspection of non optimized and delay optimized
protocols............................. 68
5.12.3.1 Death rate of nodes and rate of energy utilization . 68
5.12.3.2 Data throughput and PAR . . . . . . . . . . . . . 69
5.12.3.3 Delay ......................... 71
5.12.4 Comparative review of non optimized and energy optimized
protocols............................. 73
ix
5.12.4.1 Rate of network collapse and per instance network
energy consumption . . . . . . . . . . . . . . . . . 73
5.12.4.2 Packets loss and PAR . . . . . . . . . . . . . . . . 75
5.12.4.3 Delay ......................... 76
5.13 RE-AEDG,AEDG and ULVRP: Performance tradeoffs . . . . . . . 78
5.14 Conclusion of chapter . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6 Conclusion and future work 81
6.1 Conclusion................................ 82
6.2 Futurework............................... 83
Bibliography 84
x
List of Figures
1.1 Typesofrelays ............................. 7
4.1 Example of naming of sensor nodes in DB-EBH . . . . . . . . . . . 23
4.2 Network diagram of DB-EBH . . . . . . . . . . . . . . . . . . . . . 24
4.3 Control packet format . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.4 Flow chart for DB-EBH . . . . . . . . . . . . . . . . . . . . . . . . 27
4.5 Network lifetime in terms of dead nodes . . . . . . . . . . . . . . . 30
4.6 System energy consumption . . . . . . . . . . . . . . . . . . . . . . 31
4.7 Packets receiving and transmission burden over nodes . . . . . . . 31
4.8 Packets reached to sink . . . . . . . . . . . . . . . . . . . . . . . . . 32
5.1 Network configuration model . . . . . . . . . . . . . . . . . . . . . . 37
5.2 Slicing of transmission range of source node . . . . . . . . . . . . . 40
5.3 Cases of energy consumption model . . . . . . . . . . . . . . . . . 43
5.4 Cooperative diversity system . . . . . . . . . . . . . . . . . . . . . . 50
5.5 Sink mobility pattern . . . . . . . . . . . . . . . . . . . . . . . . . 54
5.6 Energy consumption: Feasible Region . . . . . . . . . . . . . . . . 56
5.7 Delay generation: Feasible Region . . . . . . . . . . . . . . . . . . 60
5.8 Relative examination of LVRP and U-LVRP . . . . . . . . . . . . . 65
5.9 Death rate among nodes in comparison to delay optimized protocols 67
5.10 Network energy consumption in comparison to delay optimized pro-
tocols .................................. 69
5.11 Data acceptance rate in comparison to delay optimized protocols . 70
xi
5.12 Packets approaching to the surface sink in comparison to delay op-
timizedprotocols ............................ 71
5.13 Message latency in comparison to delay optimized protocols . . . . 72
5.14 Number of alive nodes per interval in comparison to energy opti-
mizedprotocols ............................. 73
5.15 Energy consumption in comparison to energy optimized protocols . 74
5.16 PAR of networks in comparison to energy optimized protocols . . . 75
5.17 Packets drop rate in comparison to energy optimized protocols . . . 76
5.18 E2E delay in comparison to energy optimized protocols . . . . . . 77
xii
List of Tables
1 Acronyms ................................ 1
2 Nomenclature.............................. 2
2.1 Comparison of the state-of-the-art work . . . . . . . . . . . . . . . . 15
4.1 Simulation parameters for DB-EBH . . . . . . . . . . . . . . . . . . 29
5.1 Attributes of types of deployed nodes . . . . . . . . . . . . . . . . 38
5.2 Simulation constants for LVRP and ULVRP . . . . . . . . . . . . . 64
5.3 Simulation specifications for RE-AEDG . . . . . . . . . . . . . . . . 66
5.4 Performance tradeoffs . . . . . . . . . . . . . . . . . . . . . . . . . . 80
xiii
Table 1: Acronyms
WSNs Wireless sensor networks
UWSNs Underwater wireless sensor networks
AUV Autonomous underwater vehicle
OEB Optimal energy allocation for linear networks
FAF-EBRM Energy-balanced routing method based on forward-aware factor
BTM Balanced transmission mechanism
EBRP Energy-balanced routing protocol
H2−DAB Hop-by-Hop dynamic addressing based routing protocol
C-ARQ Cooperative automatic repeat request routing protocol
SRC Selective relay cooperation
DNC Dynamic node cooperation based protocols
COBRA Cooperative best relay assessment
DEADS Depth and energy aware dominating set based algorithm for coop-
erative routing along with sink mobility
AEDG An efficient data-gathering routing protocol
AEDG-Dopt Delay-optimized efficient data-gathering routing protocol for
UWSNs
AEDG-Eopt Energy-optimized efficient data-gathering routing protocol for
UWSNs
LVRP Layered voronoi scoping-based routing protocol
LVRP-Dopt Delay-optimized layered voronoi scoping-based routing protocol
LVRP-Eopt Energy-optimized layered voronoi scoping-based routing protocol
RE-AEDG Reliable AUV-aided Efficient Data Gathering Routing Protocol for
Underwater Wireless Sensor Networks
RE-AEDG-
Dopt
Delay-optimized Reliable AUV-aided Efficient Data Gathering
Routing Protocol for Underwater Wireless Sensor Networks
RE-AEDG-
Eopt
Energy-optimized Reliable AUV-aided Efficient Data Gathering
Routing Protocol for Underwater Wireless Sensor Networks
DB-EBH Depth-Based Energy-Balanced Hybrid Routing Protocol for
UWSNs
AF Amplify and forward
DF Decode and forward
MRC Maximal ratio combining technique
ERC Equal ratio combining technique
GWs Gateways
MNs Member nodes
REG Residual energy grade number
NACK Negative acknowledgement
Dim Dimensional
BER Bit error rate
PAR Packets acceptance ratio
SNR Signal to noise ratio
E2E delay End to end delay
1
Table 2: Nomenclature
Itran Acoustic intensity at transmitter
Ireceiver Acoustic intensity at receiver
RDistance between transmitter and receiver (m)
auAbsorption coefficient
fFrequency (KHz)
kSpreading factor
NsShipping noise
NwWind noise
NthThermal noise
NtTurbulence noise
SSource node
CCooperating node
DDestination node
sSlice
Es1
SEnergy consumption at S for sending data towards slice 1
Es2
SEnergy consumption at S for sending data towards slice 2
Es3
SEnergy consumption at S for sending data towards slice 3
S1
eTotal sensors belonging to first layer
nTotal number of nodes
NSet of n nodes
NUNodes deployed above to the center of the deployment field
NLNodes deployed below to the center of the deployment field
De Depth of the field
dDepth between the transmitter and surface sink
b
NSet of neighbors
N′Alive nodes
ηcPossible cooperating nodes set
ηdPossible destination nodes set
BcBest cooperation set
βWireless link quality
seSender node
2
rereceiver node
SS,D Signal sent from source and receive at destination
SS,C Signal sent from source and receive at cooperating node
SC,D Signal sent from cooperating node and receive at destination node
αChannel responses
nS,C Channel noises between source and cooperative relay
nS,D Channel noises between source and destination
xOriginal sent signal
PD
zProbability that z belongs to destination node
MAUV Length of major axis
mAUV Length of minor axis
EEnergy
i, j, z Node’s ID
hiHop of i
ρhiPredecessors of hop i
ReResidual energy
PePacket energy
Dy Delay
StSojourn time
HHolding time
PPackets generated
Ccap Channel capacity
Wnet Network width
TxR Transmission range
TTransmission of packets
3
Chapter 1
Introduction
4
1.1 Why underwater sensor networks?
Oceans and seas remained unattended for eras but then a time reaches that the
emerging nations realized the significance of the deep marines. Oceans got their
deserved economical, social, geological and ecological importance and currently the
nations are fully aware of it. In Underwater sensor networks (UWSNs) multiple
unmanned underwater vehicles are deployed to form a network. Deployed sensors
form a collaboration and helps the human race in the exploration of underwater
resources. Generally these nodes are equipped with the required sensors. They are
responsible for the observation of demanded scientific or environmental information
from the atmosphere.
1.2 Types of UWSNs and challenges
WSNs could be essentially distinguished in two types: homogeneous and hetero-
geneous. In heterogeneous network, nodes may possess different features in the
terms of installed battery power, communication and sensing range, transmission
power, node architecture, hardware complexity and much more [16]. UWSNs have
multiple operations that covers vast domain of applications but this comes with a
number of challenges that the deployed network faces inside the prescribed habitat.
Underwater atmosphere depicts a lot of differences in contrast to terrestrial envi-
ronments [1, 2] including large latencies, greater attenuation levels, higher trans-
mission energy, lower data integrity, difficult extraction of data, un expected failure
of nodes, limited availability of bandwidth, confined battery powers etc.
1.3 Applications of UWSNs
In-spite of the above listed problems scientists still pay keen attention to this field
which is an emerging domain and due to its extended applications range can’t be
neglected anymore [3, 4]. Few of them are mentioned below;
5
•Seismic monitoring : Use of UWSNs in petroleum industry such as su-
pervision of underwater oil fields and gas reservoirs.
•Data Logging : Temperature, pressure, acoustic life, water quality moni-
toring and much more.
•Military Applications : Submarines movement detection, surveillance,
navigation, tracking and a lot more.
•Disaster Management : Flood detection, revelation of the earthquake
possibility, calamity warnings etc.
1.4 Environmental constraints
It is a fact that UWSNs present a huge list of applications however, deployed
wireless nodes come across number of challenges inside the defined climate. The
very major constraint, with which the wireless nodes mostly come across, is the
limited energy resources available to them. Since the battery power attached to
a node can be of limited size and power. Moreover recharging and replacing of
batteries is also a difficult task for the networks deployed at remote areas because
at these regions human access is difficult or nearly impossible e.g. underwater oil
and gas pipeline, submarine cables and much more.
1.5 Communication channel
Due to above mentioned constraints, for the sensor networks, underwater medium
is considered to be severe geographical environment. This is due to its high associ-
ated cost and unique characteristics such as; frequency dependent attenuation and
comparatively lower data propagation speed [5].Therefore, underwater surround-
ings does not facilitate high frequency radio signals to propagate, since it provides
<100kHz of bandwidth range.
Consequently, for underwater scenarios extra low frequency radio signals (30-
300kHz) were found to be compatible. However this requires larger antennas and
6
Sink
Source node
Relay nodes
Comm. link
A B
Figure 1.1: Types of relays
higher transmission powers. So higher frequency radio signals are more immune
to the attenuations, signal distortion, end to end delays, data loss, higher BER
due to multi path fading and many more. Thus for UWSNs, acoustic channels are
considered to be more appropriate for data communications [6].
Despite of the adaptation of acoustic channels the lower data reliability is still
a biggest threat for scientists and researchers. Therefore, we propose a routing
algorithm that employs cooperative communication as a solution to tackle this
alarming complication. Since it could provide additional throughput and data
reliability gains to the network.
1.6 Relay communication
Cooperative communication is also known as, relay communication. Relays are the
nodes by which means source and destination nodes are hooked together. Relay
nodes could assist the transmission process in two ways as depicted in Fig. 1.1;
•Nodes receive the data packets from surrounding nodes and transmits that
data towards a destination node, laying at far off place. It is shown in Fig.
1.1A,
7
•or the nodes over hear the packets from a source node, regenerate its copies
and transmits them towards a destination node, laying in the transmission
range of both source and cooperating nodes [7, 8]. It is depicted in Fig. 1.1B
In cooperative routing technique, second case is employed where each in range node
plays the role of a potential relay. In cooperative models eventually the destination
receiver hears the packets from the transmitters and combines them using some
predefined diversity combining techniques e.g. Maximal ratio combining (MRC),
Equal ratio combining(ERC) and others [9, 10].
1.7 Thesis outline
Rest of the thesis is organized as follows. Chapter 2 describes the related work and
overview of counterpart schemes. Chapter 3 presents underwater acoustic chan-
nel characteristics. Chapter 4 shows our proposed scheme Depth based energy-
balanced hybrid technique, its network model, protocol operations and analysis of
our proposed model in reference to a pre-existing scheme. Chapter 5 consists of
another proposed model; improved reliability and AUV-aided efficient data gath-
ering routing protocol for UWSNs. It formulates the energy consumption model
and system model. Moreover, discuses the relay selection scheme and AUV sink
trajectory. Furthermore it includes two optimization techniques for very basic pa-
rameters of the system: end to end delays and energy consumption. At the end of
chapter performance efficiency of the proposed technique is examined in the com-
parison of two counterpart schemes. Chapter 6 concludes the chapter whereas;
references are also mentioned at the end of thesis for the interested readers.
8
Chapter 2
Literature review
9
Number of protocols have been proposed that tries to troubleshoot the energy
constraint affixed to UWSNs. Few are mentioned below:
A network model with name optimal energy allocation for linear networks-OEB
has been proposed in [11]. In this model each lower depth node accepts and retrans-
mits the data received from higher depth nodes. This retransmission mechanism is
the main cause of imbalanced energy consumption in a moored monitoring system.
Hence, in order to cater this situation, OEB combines both the multihop as well as
direct communication technique. Nodes alternatively switch their communication
modes, depending upon the comparison of energy levels of the sender node and its
neighboring node.
Energy-Balanced Routing Method Based on Forward-Aware Factor (FAF-EBRM)
is suggested in [12]. Energy balancing is done through computing the forward
transmission area for the packets flow. This protocol also defines the forward en-
ergy density and addresses the linearly random deployed network. In FAF-EBRM,
the next-hop node is selected according to the awareness of link weight and their
forward energy density.
Jiabao Cao and few others have proposed a balance transmission mechanism
in [13]. It divides the data routing process into two segments. In the first step it
creates a tree of nodes which is actually a path, a packet follows in data transmis-
sion process. This path is created on the basis of optimum transmission distance.
On the second stage data routing algorithm is designed. Multihop and single hop
both methods of communication are employed to balance the energy consump-
tion. Energy gradation concept is also used in order to decide the transmission
mode of nodes. Simulations and results prove the reliability of the protocol in the
enhancement of network lifetime.
In [4A] authors proposed an energy balanced protocol in which three potential
fields are defined using depth, energy density and residual energy. This protocol
divides the deployment area into three different layers. This basic approach helps
in the movement of packets from dense energy area hence saving the lower energy
nodes to indulge in more transmissions and resulting in dissipation of energy. This
10
paper also deals with the loop creation problem, that happens due to the priority
given to higher energy nodes.
(H2-DAB) is a routing protocol for UWSNs mentioned in [15] has been dis-
cussed. This protocol employs the hop by hop data communication technique. It
takes the advantage of multiple sink deployment in the network. It also claims the
cost effectiveness as it does not requires special sensors hardware modules for loca-
tion or depth tracking of randomly deployed nodes. Through experimentation and
simulations this protocol shows the lower message latency, better packet delivery
ratio and reduced energy consumption of the network.
Aforementioned are the few protocols taking part in balancing the energy con-
sumption for an emerging and critical domain of wireless sensor networks. Besides
that researchers and scientists are also menaced by the nature of acoustic chan-
nel attributes. In order to overcome the shortcomings of channel, WSNs have
used diversity in terms of frequency, time or space. Cooperative communication
emerged as an alternative technique of space diversity. It takes the advantage of
the key factor which was initially considered as an computational overhead for
the nodes: Each node over hears the un destined packets from its own vicinity.
Hence, these nodes provides the antenna diversity to the system by relaying those
unrequested received packets towards their destination. This technique yields var-
ious paths for a single packet circulation and each route constitutes nonidentical
channel attributes. Therefore produces independent fading effects.
Initially it was formulated for the sender node to resend the packets after regular
intervals until it receives some acknowledgment message from receiver side. This
technique generally asks for several retransmissions from the same node. This is
due to the fact that packet passes out of the channel that posses almost similar
characteristics every time. Hence a better version of ARQ protocol was presented
in [17]. This scheme suggests that receiver node should ask for the retransmission
from the cooperating nodes except of the sender node itself. Hence, network life
time is substantially improved using this method.
11
Dynamic node cooperation in UWSNs is also presented in [18]. Basic ARQ
scheme has been refined in this protocol. This mechanism never asks retransmis-
sions from the same node whose data was un-decoded formerly. The receiver asks
for a retransmission from the node that has overheard the same packet and got
better channel conditions. In this protocol the relay nodes are preferred on the
basis of link quality and SNR. Authors implemented the protocol on real systems
and performed the lake tests. They demonstrated the effectiveness of the offered
system this way.
After the basic implementation of cooperative communication another chal-
lenge is emerged. That states as: which neighboring node would be considered
as a relay node? Different papers present unique relay node selection standards.
Work at [19] took the channel statistical information and proposed a best relay
selection criterion. It takes into account both the incurred propagation delay and
transmission time. In this paper author also formulated an optimum one way
packet transmission (OPT) time. This protocol is proved to be more sensitive
towards the source to relay channel as compare to relay to destination channel. It
also proves that the shorter OPT time will improves the throughput rate of the
network.
In [20] authors state a unique cooperating nodes selection norm. It is based
on the propagation delay and then the SNR factor is also incorporated to produce
the optimal results. This scheme considerably minimizes the amount of required
retransmissions among the nodes. Thus it is deduced that the selection of cooper-
ating node may leads to efficacious energy consumption of the network.
Cooperative communication was emerged as excessive energy consumption tech-
nique in contrast to non cooperative models; which is a complication. Thus a pro-
tocol was suggested in [21] that compensates the exceptional amount of consumed
energy by incorporating the sink mobility inside the considered field. AUV follows
an elliptical trajectory and takes sojourn stops after regular intervals of time. This
way the intermediate distance between the nodes and surface sink is minimized.
Consequently the overall energy remains conserved.
12
Multiple techniques for relay communication have been discussed, compared
and implemented till now [22, 23]. Both mechanisms have their own strengths and
weaknesses. However our proposed model incorporates the Decode and Forward
(DF) mechanism as a relay communication technique.
2.1 Overview of counterpart schemes
For DB-EBH the considered counterpart scheme is Hop by hop. Whereas the
framework of RE-AEDG is mostly related to AEDG and LVRP as discussed in
[24, 25] respectively. Hop by hop scheme is selected because our proposed DB-
EBH is a combination of multi hoping and single hoping. Moreover it is intended
to portray the effectiveness of hybridization over a simplistic non hybrid tech-
nique. Multihoping deposits exceptionally intensified packets routing burden over
the lower depth nodes causing network partition and prior battery power deple-
tion among nodes. Furthermore AEDG and LVRP are chosen to place in proximity
because of their conceptual relevancy with RE-AEDG.
The key point to insight here is that these protocols conserves the network
energy but does not contributes towards higher level data authenticity to the re-
searchers. Hence a protocol was needed that should be well organized, cost ef-
fectual and handles that preserved energy to upgrade the data validity factor for
the exalted outcomes. Therefore as a consequence of these mentioned features we
propose a protocol that is an enhanced version of scheme AEDG.
RE-AEDG is focused to magnify the system reliability by magnifying the
throughput rate and successful packets delivery ratio at the surface sink. Hence
in order to attain the aforementioned prototype, a strategy known as cooperative
communication is applied. This scheme is appropriate because it substantially
boosts the spectrum utilization without placing any extra cost liability over the
system except higher energy consumption.
Conservation of energy is the distinguishing factor of AEDG. Thus the design
objective of our technique is the effective utilization of that uphold energy to
intensify the network’s operational efficiency.
13
The analogy between the aforesaid schemes (AEDG and LVRP) in the reference
of our proposed technique (RE-AEDG), is examined underneath.
•Both involves the multihoping routing method accompanied with the surface
and Mobile sink(s).
•Focuses on the balanced as well as the reduced energy consumption of the
network.
•Minimizes the conventional protocol overhead for routing management by
associating GWs or anchor nodes.
•Does not caters the reliability of data, which is a critical issue to mull over
in harsh underwater environment.
•Pays the cost of higher level delay induction.
•AEDG associates single sink while LVRP incorporates several sinks.
•AEDG follows the predefined orbit whereas LVRP includes the random mo-
tion for AUVs.
•For AEDG network is heterogeneous while for LVRP network is homoge-
neous.
•AEDG collaborates with numerous GWs per instance but LVRP associates
single anchor node per AUV.
14
Table 2.1: Comparison of the state-of-the-art work
Technique/
Ref Objective(s) Achievement(s) Deficiency(ies) Field /Architec-
ture
OEB
[1A]
i- Per node balanced
energy consumption
ii- Omit energy hole
problem
i- Better stability period
ii- Network Lifetime
iii- Lowered packet
forwarding overhead
i- Extended delays in contrast
to hop by hop
ii- Data integrity is
unconsidered
i- 2Dim-Ocean
mooring
model
FAF-
EBRP
[2A]
i- Energy balanced
routing
ii- Robust topology
iii- Fault tolerance
i- More last surviving nodes
ii- Packets reception ratio
iii- Balancing of energy
iv- Higher QoS factor
i- End to end delay
ii- Frequent forward trans-
-mission area calculation
iv- Computational overload
i- 2Dim-UWSN
ii- Cluster based
rectangular model
architecture
BTM
[3A]
i-Balancing energy
ii- Optimized energy
level calculation
iii- Energy hole
reduction
i- Optimal path selection
ii- Prolonged network’s
operational duration
iii- Outperforms
scalability test
i- Control packet overhead
ii- Routing loops creation
iii- Message latency
i- 2Dim-UWSN
ii- Uniform
random circular
deployment
EBRP
[4A]
i- Protect low energy
nodes
ii- Extended coverage
ratio
iii- Stability period
enhancement
i- Loop detection and
elimination
ii- Robust against noise
iii- Maintains consistent
performance
iv- Prolong network lifetime
i- Improved end to end delays
due to non consideration
of shortest path routing
ii- Protocol overhead
iii- Loops creation
iv- Update packets exchange
i- 2Dim-UWSN
ii- Homogeneous
flat architecture
H2DAB
[5A]
i- Time efficient
routing
ii- Energy efficiency
iii- Scalability
i- Localization free protocol
ii- Higher data deliver ratio
iii- Not maintains
complex routing tables
iv- Energy conservation
i- Imbalanced energy usage
ii- Higher packets
re-forwarding burden
iii- Message latency
i- 3Dim-UWSN
ii- Anchored nodes
iii- Multiple sink
architecture
C- ARQ
[12]
i- Provide alternative
routing paths
ii- Retransmission of
erroneous data
i- Exploits broadcast nature
of transmissions
ii- Throughput efficiency
i- Energy consumption
ii- Higher E2E delay
iii- Packets acceptance
iv- Network partition problem
i- 2Dim-UWSN
ii- Two terminal
cooperative relay
model
SRC
DNC
[13]
i- Best relay node
selection
ii- Data authenticity
iii- Less retransmissions
i- Enhanced data integrity
ii- Improved packet
reception rate
i- Non efficient energy usage
ii- Higher E2E delays
iii- Higher protocol overhead
i- 2Dim-UWSN
ii- Random
deployment
COBRA
[14]
i- Best relay selection
ii- Overcome frequency
selective fading
iii- Minimization of one
way packet trans-
-mission (OPT) time
i- Throughput improvement
ii- Reduced delay if relay
is selected for minimized
OPT time
iii- Less packets collision
iv- Lowered outages
i- E2E delay dominates
if packet size is small
ii- Increased energy
consumption
iii- Network partition
i- 2Dim-UWSN
ii- Random
deployment
[15]
i- Relay selection based
on propagation delay
and SNR
ii- Data authenticity
i- Outperforms other
techniques in
terms of delay induction
ii- Improved bit error rate
i- Lowered stability period
ii- Reduced network lifetime
iii- Extended energy utilization
i- 2Dim-UWSN
ii- Random
deployment
DEADS
[16]
i- Data reliability
ii- Throughput
iii- Energy efficiency
iv- Adaptive routing
i- PAR
ii- Reduced packet drop
iii- Reduction in
energy consumption
iv- Optimized dth
i- Extended E2E delays
ii- Control and processing
overhead
iii- Stability period
compromised
i- 3Dim-UWSN
ii- Randomly
deployed
iii- Two sink
mobilities
AEDG
[19]
i- Reduced as well as
balanced energy
consumption
ii- Efficient data
gathering
iii- To cut the internodes’
communication distance
i- Improved stability period
ii- Prolonged network
lifetime
iii- Comparatively reduced
packet drop
i- No significant raise in data
authenticity
ii- Message latency
iii- Control packets overhead
i- 2Dim-UWSN
ii- Non homo-
-geneous random
deployment
iii- Single AUV
LVRP
[20]
i- Efficient distributed
data routing
ii- Reduced routing
update frequency
iii- Energy conservation
i- Alleviated energy hole
problem
ii- Improved packets
delivery ratio
iii- Less forwarding
overhead
i- Path selection overhead due
to random movement of AUV
ii- AUVs leave field un-
-attended and penetrate in
one another voronoi’s scope
iii- Long communication paths
i- 2Dim-UWSN
ii- Uniform random
circular
deployment
iii- Multiple AUVs
15
Chapter 3
UWSNs: Communication models
16
While propagating through underwater medium energy density of the signal
decays due to water absorption and properties of acoustics channel. Therefore
precise models are required to model the underwater communication channels. We
present few models here firstly the attenuation model is stated. It presents the
attenuation factor between transceivers. Attenuation is directly affected by the
transmission path length. Afterwards the transmission losses and the underwater
noise features are modeled.
3.0.1 Attenuation model
Underwater acoustic signals experiences power level degradation. Hence reduced
signal strength is the source of attenuation implantation in the message signal.
Attenuation is calculated for distance in meters and frequency in KHz and is given
as follows:
A(R, f ) = Rkau(f)R,(3.0.1)
au(f) is measured in dB/km where as k differs with geometry of propagation. It is
2,1 and 1.5 for spherical, cylindrical and practical spreading respectively [31–33].
Absorption coefficient can be presented using Thorp’s empirical formula mentioned
as under:
10loga(f) = 0.11 f2
1 + f2+ 22 f2
4100 + f2+ 2.75 ×10−4f2+ 0.003,(3.0.2)
above written equation depicts that absorption coefficient is directly related to the
value of frequency.
3.0.2 Transmission loss model: Deep and shallow water
Transmission loss also known as propagation loss defines the decrease in energy
level of signal as wave propagates through the specified medium. For shallow water
the propagating acoustics signals experiences the sea bed and water surface both so
it behaves as a cylindrical wave and cylindrical spreading can be observed [34, 35].
Transmission loss is defined as the ratio of the signal’s intensity calculated at the
17
transmitter to the intensity measured at the receiver. If the receiver lies at the
distance of 1m then the transmission loss can be represented as:
Itran
Ireceiver
(3.0.3)
Transmission loss caused by cylindrical spreading along with attenuation factor is
represented in equation written below:
T L = 10log(R) + au(R)×10−3(3.0.4)
Whereas for deep sea, environment is considered to be uniform causing the spheri-
cal wave spreading. Therefore transmission loss for deep sea would be represented
as:
T L = 20log(R) + au(R)×10−3,(3.0.5)
3.0.3 Noise model
In the underwater medium sound propagates with a speed of 1500m/s. This de-
picts that the radio signals can propagate with five times faster speed as compare
to acoustic signals in underwater environment. Environmental noise is due to
three main factors that are scattering, absorption and spreading loss. These losses
are mainly caused due to the unsteady movement of water, ships motion for the
sake of tourism or trade, military operations etc. Although the environmental
temperature is a natural process but it is always not favorable for surroundings.
Specifically in the under water scenarios it causes higher noise levels. Therefore,
while modeling the attenuation for underwater environments all these factors are
incorporated [36–38] and shown as below:
10logNs(f) = 40 + 20(s−0.5) + 26log(f)−60log(f+ 0.03),(3.0.6)
10logNω(f) = 50 + 7.5ω1/2+ 20log(f)−40log(f+ 0.4),(3.0.7)
18
10logNth(f) = −15 + 20log(f),(3.0.8)
Total noise power spectral density appearing on the signal can be represented as:
NL(f) = Nt(f) + Ns(f) + Nω(f) + Nth(f) (3.0.9)
It is also to note that all the four noise factors are directly dependent over the fre-
quency, that’s why fluctuation in frequency effects the communication undergoing
the environment. In [42] authors worked on the channel characteristics involving
the channel attenuation model. They also consolidate the information signal due
to the movement of receiver nodes known as Doppler effects. In their implementa-
tion they calculates the optimum frequencies that acquires better SNR, providing
data communication possible at relatively larger distances.
19
Chapter 4
DB-EBH: Depth-Based
Energy-Balanced Hybrid Routing
Protocol for UWSNs
20
4.1 Summary of chapter
Underwater wireless sensor networks (UWSNs) are meant to be deployed at the ar-
eas that need to be monitored continuously without the human assistance. There-
fore, these networks are expected to stay operational for a longer period of time.
However, sensor nodes in these networks are equipped with limited energy (e.g.,
battery) resources. Moreover, uneven energy consumption is one of the biggest
challenges in UWSN because it leads to creation of energy holes and ultimately
shorten network lifetime. This invites UWSN designers to introduce protocols that
can minimize and balance energy consumption of nodes. This section presents DB-
EBH, a Depth-Based Energy-Balanced Hybrid routing protocol for UWSNs. Like
EBH, DB-EBH is a hybrid approach which is based on direct and multihop com-
munication. However, DB-EBH considers linear random deployment of nodes. It
selects a priority neighbor node for data forwarding on the basis of its depth from
the sink. Simulation results validate the performance of DB-EBH in terms of
energy efficiency, network lifetime and throughput.
4.2 Motivation
Most of the work done in UWSNs is focused to cut short the overall consumption of
the network. Whereas, EBH [26] takes the energy balancing for individual nodes.
It highly affects the network by improving the network stability time. However,
EBH is defined by considering a linearly deployed network with fixed inter-node
distance. This deployment state may not be feasible sometimes. Generally sci-
entists encounters the underwater environments which are unkind enough to not
allow the fixed network deployment. Hence, in these atmospheric conditions it is
feasible to deploy the network randomly. Therefore, our protocol assumes a lin-
early random deployed network with unequal intermediate distances among the
nodes.
21
4.3 Protocol operation
Our suggested protocol DB-EBH: employs the basic concept of EBH. It divides
initial energy of nodes into number of smaller chunks known as energy grades.
One small chunk is known as one energy gradation. Our proposed protocol works
by alternatively switching between both modes of communication depending upon
how faster a node dissipates its one grade of energy. In order to find the optimum
number of energy levels a method has been discussed in [26] and is followed in our
technique. Optimum calculation of the energy levels is required because higher
number of energy grades can result in smaller chunks of energy and therefore can
cause the transmission modes to switch rapidly. Rapid switching among the modes
of communication can leads to a larger burden of control packets forwarding over
the network. Contrary to this smaller number of energy grades leads towards
the larger energy chunks that result in the transmission mode to be in multi hop
state maximally causing imbalanced energy consumption. This happens because
multihop communication dissipates smaller amount of energy as compare to the
direct transmissions.
Considering the scenario of two consecutive nodes: The higher depth neigh-
boring node is known as the successor node whereas; the lower depth neighboring
node is called as the predecessor node. Fig. 4.1 depicts the basic model to further
elaborate the concept. Whereas step by step process of the packet’s transmission
among nodes is discussed underneath:
•Each node keeps an eye over its residual energy grade (REG).
•Once the energy grade of a node falls, it informs its predecessor node by
releasing a control packet towards it.
•This packet contains the node’s id and its REG number.
•Upon reception, predecessor node checks if its own REG is higher than that
of its successor node.
22
N
N-1
N-2
N-3
N-4
N-5
N is the successor node of N-1
N-1 is the predecessor node of N
N-3 is the successor node of N-4 & N-5
N-4 & N-5 has no predecessor node
Figure 4.1: Example of naming of sensor nodes in DB-EBH
•If yes, it acknowledges the successor node. Furthermore takes the packets
from its predecessor neighboring nodes and switches towards direct trans-
mission mode.
•Consequently, the predecessor node reduces the transmission load over its
successor node which now needs to multihop its own sensed data packets
only.
•Currently the predecessor node would be dissipating higher energy as com-
pared to its successor node.
•Thus a time reaches when their REGs get similar again and the predecessor
node re switches to the multihop mode, and the process continues.
23
Multi-hop
Single-hop
Sensor node
Base station
Sink F
C
A
B
D
E
Figure 4.2: Network diagram of DB-EBH
4.4 DB-EBH: Network model
In order to find out the depth of every sensor (in comparison with the surface sink),
our proposed model assumes a localization free protocol that picks the nodes being
equipped with the inexpensive depth sensors.It also considers that once a packet
reaches the surface of the water, it is assumed to be successfully received by the
sink located at the water top. Whereas if a node gets no neighbor in its transmis-
sion range then all the packets would be forwarded directly towards the sink.
The main concept of the working principle of Depth based EBH is depicted in
Fig. 4.2.
•A bunch of randomly deployed underwater nodes and their inter communi-
cation is being considered.
•All the nodes are in multihop state except the node Dwhich is in direct
transmission mode. It is forwarding its own sensed packet as well as the
24
packets from the predecessor nodes Band C.
•This way node Ehave reduced burden of packets forwarding because ex-
cept of forwarding packets from its predecessor node Dnow it will have to
multihop its own sensed packets only.
•Hence after few rounds of transmission, the direct forwarding state of node
Dwill leads to the higher energy dissipation rate. This results in the rapid
fall in energy gradation.
•At that instant the energy grades of both the nodes Dand node Ewould be
similar.
•Therefore, node Dwill ping its successor node. Both of them recreate a link
in between and resume the inter communications.
•In the mean time node Dreleases a control packet towards its predeces-
sors. It demands them to switch their transmission mode towards the direct
transmission state and the process continues.
For the sake of the initialization of the network, the communication process starts
by broadcasting of hello packets by all the nodes. Nodes searches for the the eligible
neighbors in its transmission range depending upon the depth information shared
by them. Nodes kept on forwarding those control packets after regular intervals
of time to update the surrounding nodes about their current state of energy and
depth. Format of hello packets is shown in Fig. 4.3.
Node ID Residual Energy Grade No.
Figure 4.3: Control packet format
25
4.5 Priority node selection
For DB-EBH the energy gradation concept is same as being followed in EBH.
However in this technique the selection of neighboring nodes and then the priority
neighbor depends upon the following conditions:
•Receiving nodes must be at lower depth as compare to the forwarder node.
•Receiving node must lays in the transmission range of the forwarding node.
•Finally, the most important check is upon the depths of the nodes present in
the neighboring list of the sender node. A neighboring node that is present
at the minimum depth in comparison with the other neighboring nodes, is
considered as the highest priority neighbor.
To clarify the working concept of DB-EBH, a flow diagram is shown in Fig. 4.4
that shows the basic steps being followed in protocol’s designing process.
4.6 DB-EBH: Operational flow
Every node broadcasts its hello packet in its transmission range. Other nodes
receive this packet, compare its depth with its own depth and decide the eligible
neighbors Eligible neighboring node’s IDs are stored in a qualified forwarder list.
Qualified forwarder list is further sorted on the basis of the minimum depth. The
node at the first index of list will remain the priority neighbor for its lifetime. After
the death of the priority neighbor the node placed second in the list will take its
position and so on.
4.7 Simulation results and discussion: DB-EBH
and HBH
In this section we evaluate the performance efficiency of our proposed protocol. For
performance evaluation of our proposed technique DB-EBH, we simulated another
existing technique known as Hop by Hop (HBH). HBH incorporates multihop
26
Start
Di < Dj
Di – Dj >=
TxRange
yes
No Packet drop
Store the node’s ID
in the List
Calculate the
energy grade
Eo/m of the node
yes
Sustain the
communication
mode
One grade of energy
falls yes
Release a control
packet towards
upstream node
REG of Predecessor
>
REG of Successor
Yes
Switch towards
direct transmission
mode and
acknowledge the
predecessor node
No
Continue the
existing mode of
transmission
Initialize all the
nodes with
multihop
communication
scheme
No
Direct
transmission
mode
Transmit packet
Yes
yes
Control packet
recieved
Continue the
existing mode of
transmission
No
End
Transmit the
received control
packet towards its
successor node
No
No Yes
J == N
Yes
Sort the
neighboring list
Multihop
communication
mode
No
Hello packet
received
Figure 4.4: Flow chart for DB-EBH
27
routing whereas DB-EBH is a merger of two: multihop and direct transmission
techniques. For a fair comparison two randomly deployed networks are considered.
We suppose a 500meters ×500 meters field with randomly deployed nodes. A
sink is considered which is assumed to be floating over the water surface. Death
threshold is also defined which explains the minimum energy after which node is
considered to be dead. Packets reaching to the sink tell the the number of packets
reaching at the water surface successfully.
4.7.1 Definition of network performance parameters
4.7.1.1 Network lifetime
It is the time after which network is considered to be un-operational. In our case
we taken it till the death of half the number of network nodes. It is calculated in
seconds.
4.7.1.2 Stability time
It is the measured in seconds. It is the interval measures before the death of the
first node of network.
4.7.1.3 Packets to sink
Packets reaching to the sink tells the number of packets reaching at the water
surface successfully. It is calculated in reference to the number of packets being
forwarded to the sink.
4.7.1.4 Packets to nodes
It tells the packet forwarding burden nodes caries while transmitting the packets
of predecessor nodes.
4.7.1.5 Residual energy
It explains the overall network remaining energy. It is measured in Joules (J).
28
Table 4.1: Simulation parameters for DB-EBH
Simulation Parameters Values
Initial Energy of a node Eo 0.02J
Number of Energy grades 1.623
Death criteria Ee Eo*106nJ
Network dimensions 500 m2
Default frequency of acoustic modem 25 kHz
Number of wireless sensor nodes 260
Transmission range for nodes in multihop 80 m
4.7.2 Results and discussions
In Fig. 4.5 the network life time of Hop By Hop and DB-EBH is compared. The
clear difference of their network life time can be visualized easily. This is because
in HBH data routing technique the lower depth nodes are more burdened. This is
due to the fact that in DB-EBH few nodes are switched to direct transmission state
resulting in lowered burden of retransmission among the nodes. Consequently, an
improved network lifetime for DB-EBH is achieved. However, the lowered network
lifetime and stability period of HBH suggests that in this technique the minimum
depth nodes carry the maximum data retransmission load among themselves. This
leads to the rapid and imbalanced energy dissipation in the network. It causes the
energy hole problem for nodes and also it lowers the operational efficiency of the
network. In our mechanism hybridization of direct communication technique with
multihop technique counterbalances the routing overhead caused due to multihop-
ing and improves the network performance.
Fig. 4.6 shows the comparison of per round energy consumption of the two
protocols. Initially the energy consumption of HBH is nearly the same as the DB-
EBH but later on both protocols show clear margin in the energy consumption of
the network per round. This difference is majorly because of the higher number
of nodes involved in the data transmission process in HBH technique. More the
number of hops involved in data transmission phase more would be the utilization
29
26 28 30 32 34 36 38 40
0
20
40
60
80
100
120
Number of rounds
Number of dead nodes
DB−EBH
HBH
Figure 4.5: Network lifetime in terms of dead nodes
of computational powers at the individual nodes. DB-EBH involves direct trans-
mission too, due to which few nodes with comparatively higher residual energy
bear the more network burden and communicates the data directly towards the
sink.
Fig. 4.7 shows the number of packets received by the nodes from their neighbor-
ing nodes per round. HBH has higher transmission burden over the nodes because
a packet reaches to the sink via intermediate nodes, hence each node has to for-
ward larger amount of data in HBH as compare to DB-EBH. DB-EBH balances
this data forwarding burden over the nodes by employing the direct transmission
technique which comes alternatively with the multihop technique. Therefore, DB-
EBH proves to consume lesser amount of energy per round as well as it also extends
the network stability time.
Fig. 4.8 shows the throughput rate of the two systems. For DB-EBH if a
node encounters with a no neighbor case, it starts forwarding the packets directly
towards the sink as the nodes in DB-EBH are also assigned the capacity of the
direct transmission when needed. Whereas for HBH protocol, if a node gets no
30
Number of rounds ×104
0 2 4 6 8 10
Residual energy of Nodes
0
1
2
3
4
5
6
DB-EBH
HBH
Figure 4.6: System energy consumption
10 12 14 16 18 20
0
500
1000
1500
2000
2500
3000
3500
4000
Number of rounds
Packets receiving burden over nodes
DB−EBH
HBH
Figure 4.7: Packets receiving and transmission burden over nodes
31
Number of rounds ×104
0 2 4 6 8 10
Packets received at Sink
0
50
100
150
200
250
300
DB-EBH
HBH
Figure 4.8: Packets reached to sink
neighbor in its transmission range then all the packets are considered as the packet
loss or non-successful transmission of packets. Hence DB-EBH shows clearly higher
ratio of data packets reached successfully at the sink as compare to HBH. Moreover
the queue of nodes remains occupied much in HBH case. This is because each
node is responsible to communicate the information packets of its higher depth
neighboring nodes. Overflow of nodes buffers causes the enhancement in packets
drop for HBH case in contrast to DB-EBH case.
4.8 DB-EBH: An inspection
This chapter explains a routing protocol that exhibits multiple strengths and de-
ficiencies. Hybridization of two techniques provides enhancement in terms of net-
work lifetime and the stability period. However overall data reliability factor re-
mains unaccounted in this work. Since at central unit the acquired information
is critically analyzed and further employed to take critical decisions. Therefore,
authentic information is an intense attraction for the researchers.
32
4.9 Conclusion of chapter
Energy is the main constraint of wireless sensor networks. Less energy consump-
tion by the WSNs cannot solve the problem as a whole, as the uneven energy
consumption of nodes also creates problems by producing the energy holes which
results in network partition. Hence this can interrupt the data flow in network.
DB-EBH is the enhanced version of EBH protocol that contributes towards the
balancing of energy consumption of the nodes. As well as in comparison of HBH,
DB-EBH also improves the lifetime of a network, lowers the overall energy con-
sumption of the system due to its hybrid communication technique, and it has
also been proved that overall data transmission burden over the nodes is highly
reduced which is clearly shown in the simulation results.
33
Chapter 5
RE-AEDG: Reliable AUV-aided
Efficient Data Gathering Routing
Protocol for Underwater Wireless
Sensor Networks
34
5.1 Summary of chapter
As mentioned earlier authentic data arrival at the surface sink is an extremely
delicate element to cater in the harsh environments. Therefore, in this section we
present a routing technique peculiarly targeting the data sensitive applications. To
enhance the data reliability cooperative communication is employed. Cooperative
communication has been exploited significantly at MAC and physical layer. How-
ever its exploration at the network layer is still required. It is based on a simplest
three to four terminal communication channel. However, cooperative communica-
tion emerged as a faster battery power dissipation mechanism leads to ill-favored
state known as energy hole problem. To circumvent this issue we took a hetero-
geneous network aided by an AUV: following a predefined trajectory. Non casual
nodes are deployed along side of AUV track. They associate ordinary nodes with
the AUV and are accountable for reduced protocol overhead. Furthermore de-
lay and energy optimized models are also suggested for our protocol. Lastly the
simulation results are demonstrated in the contrast to two other pre existing tech-
niques. The discussion of advances achieved by RE-AEDG and analogue schemes
along with the cost being paid is also investigated in detail.
5.2 Motivation
Acoustics channel response is highly fluctuating. Propagating signal typically fol-
lows multipath in-order to reach the destination. Reflection of the signal degrades
certain frequencies at the receiver causing signal attenuations. These attenuations
are purely frequency dependant therefore known as frequency selective fading. In
our work we are going to propose a cooperative communication model for frequency
selective fading channels. However cooperative communication phenomenon offers
tradeoff between the data integrity and network energy consumption. Therefore
for RE-AEDG: the cooperative communication technique is fused with an acoustic
35
underwater vehicle (AUV) following a predefined trajectory. AUV is employed be-
cause of its significance in efficient data gathering phase. It eliminates the connec-
tion in between the source and destination nodes and compensates the additional
amount of energy that is generally being consumed due to cooperation mecha-
nism. Our proposed cooperative transmission scheme involves a relay selection
mechanism too.
Here we enlist our supposed scenario for utilizing the UWSNs.
•Research ships carry UWSNs to a focused area.
•Divers deploy UWSNs to the targeted region.
•Stationary nodes are responsible for required data extraction.
•Densely deployed network provides sufficient amount of cooperating nodes
to the sender.
•Acoustic underwater vehicle (AUV) follows the predefined trajectory, receives
the information packets and relays them towards surface sink.
•Surface sink communicates the observed information with some terrestrial
central unit or satellite station.
•At central unit this collected information is used further by the scientists and
researchers to take some decisions accordingly.
5.3 System model
In this section we briefly discuss the type of field selected, the pattern of nodes
deployment and the trajectory being followed by AUV. The working principles and
communication rules have been also elaborated for the sake of better understanding
of reader. Fig. 5.1 depicts the basic view of the considered environment and shows
the node’s deployment layout. In the model under our examination, nodes are
randomly deployed in a 2D zone and can be divided into three distinct categories.
36
This classification is based upon few features discussed in the Table. 5.1 and named
as underneath:
•Member nodes (MNs),
•Gateway nodes (GNs),
•Autonomous underwater vehicle (AUV)
MNs
GWs
Sink
AUV
Comm. link
Figure 5.1: Network configuration model
Considered network involves following operations undergoing at a time:
•MNs and GWs are sensing and communicating the required information from
the surroundings.
•Nodes are measuring their residual energies and informing their predecessor
nodes through a 1 bit control packet.
37
Table 5.1: Attributes of types of deployed nodes
Attributes MNs GWs AUV
Modem Acoustics Acoustics Acoustics, Radio
Energy Lesser Greater Not a constraint
Quality Larger Smaller One
Interacts with MNs, GWs MNs, GWs, AUV GWs, Sink
Info. sensing Yes Yes No
Device mobility No No Yes
•Nodes are updating their neighboring list on the basis of received energy
updates and the number of neighbors alive.
•Nodes reserve their rights to accept and discard a packet on the basis of BER
and to ask for two resents of a packet from the sender node (may be source
or cooperating node)
Furthermore our system divides the transmission range of nodes in three slices.
Nodes belonging to each slice perform unique functionalities in the reference of
source node. Nodes belonging to same slice are the Sibling nodes.
Rule:
Sibling nodes are unable to communicate data to each other.
This rule is convenient in a way that it forces an appropriate transmission
distance in between the communicating nodes and therefore reduces the number
of hops in between the source and destination nodes. However, this rule also
causes a drawback which is reduction in number of eligible neighbors for source
node. This side effect is highly undesirable during later simulation time, when
most of the network nodes are dead. To avoid this scenario an intermediate rule
was adopted that provides the solution by dividing the 1st slice further into two.
Whereas in order to yield improved performance 1st slice is further portioned into
two layers. Layering is conducted such that the region lies in between the depth of
source node [d(S)] and [thv+d(S)] makes the 1st layer and the rest of the region
of the 1st slice makes the 2nd layer. Source node itself lies in the 1st layer. Nodes
38
residing in the 2nd layer, 2nd and 3rd would be the eligible neighbor for source node.
5.4 RE-AEDG: The proposed scheme
In our proposed technique we consider the set of nodes deployed in the considered
network. Whereas Nconsists of total nnumber of sensor nodes. Nodes set can be
mathematically written as follows:
N={1,2,3,4.....n},(5.4.1)
Considering the deployment scenario Nis divided into two subsets such that:
N=NU+NL,(5.4.2)
and they are represented as:
NU⊆N and NL⊆N, (5.4.3)
whereas,
NU=
m
X
i=1
i , m ⊆N, (5.4.4)
NL=
q
X
i=1
i , q ⊆N, (5.4.5)
Consider a deployed node i that may belongs to NUor NLdepending upon the
following conditions:
i ǫ
NUif d(i)<1
2(De)
NLif d(i)>1
2(De)
.(5.4.6)
It is considered that initially all the deployed nodes are unaware of their surround-
ings, their neighboring nodes and their relative depths. Therefore for the sake of
handshaking, nodes broadcasts a control packet in their surroundings. Control
39
S
C2
C1
D1
D2D3
C3
_S
_D
_C
Figure 5.2: Slicing of transmission range of source node
packet includes the node’s ID, residual energy and depth information. Surround-
ing nodes of source node receives the packet and applies eligible Neighborhood
Criterion to these nodes. The record of nodes that fulfills the below mentioned
rules are maintained in a list known as neighboring list.
Rule:
j ǫ
c
Niif i ǫ NUand d(i)< d(j)
c
Niif i ǫ NLand d(i)> d(j)
(5.4.7)
, Above written rule states that jbelongs to the neighboring list of iif, any one
of the below mentioned two conditions is fulfilled. Conditions are stated as:
•ibelongs to the upper node’s list and depth of iis lesser than the depth of
j,
•or ibelongs to the lower node’s list and depth of iis higher than the depth
of j
For further understanding of network model consider Fig. 5.2 where the transmis-
sion range of source node can be further divided in three slices. 1st slice contains
source node itself and possible cooperating nodes whereas, 2nd slice also contains
the possible cooperating nodes. So the set of possible cooperating nodes can be
stated as:
ηc= 2ndlayer + 2ndslice (5.4.8)
40
whereas; the 3rd slice contains ηdonly.
Hence, b
Ncan be further divided into two subsets, 1st one is the set of possible
cooperating nodes and the other is the possible destination nodes set. This is
equated as follows:
c
Ni=ηc+ηd,(5.4.9)
Now consider a node jthat may lies inside the possible cooperating node or possible
destination node category depending upon the following conditions:
j ǫ
ηcif dα
th < dj−ds< dβ
th
ηdif dj−ds> dβ
th
,(5.4.10)
whereas; in the aforementioned equation dα
th and dβ
th are the real positive num-
bers. They are changing dynamically and are dependent upon the nodes density.
Therefore two cases could be possible here:
•Nodes density greater than the marginal value,
•or nodes density drops to the marginal value.
In the first case the difference between the two thresholds is kept low so that lower
number of nodes fulfills the eligible Neighborhood Criterion. However, with the
passage of time nodes start depleting their energies. At that instant keeping lower
difference among thresholds would not be a good norm. This is because it may
leads to a state that source node is unable to find sufficient neighbors around, re-
mains unattended and results in minor assistance by the cooperating nodes. This
would significantly influence our focused objective which is the reliability maxi-
mization. Therefore, with increased sparseness of network the difference between
the thresholds increase and leads to the probability that good quantity of nodes
still remains qualified for data corporation purpose. Therefore, our proposed model
is adaptive in this aspect. Mathematically it can be shown as below and ωand δ
41
are the two real numbers in the equation mentioned below:
dα
th −dβ
th > δ if N′<1/3 (N)
Ω< dα
th −dβ
th < δ if N′>1/3 (N)
(5.4.11)
Our proposed model also involves GWs, these are the casual nodes with same
physical features as that of MNs. The only difference they reserve is of the higher
installed initial battery power. The need of high level battery power was perceived
because it is generally observed that the chain heads got the maximum data trans-
mission responsibility. This burden is mainly of two factors.
•Our protocol says no to data aggregation due to its less prone nature towards
data integrity.
•Data communication technique being followed in our mechanism is multi
Hoping. As our stratagem does not supports aggregation, so the network
pays it off by the increase in packet re-forwarding burden over the nodes. In
our technique each node transmits its own data packet as well as the data
packets being generated by the predecessor nodes.
This data forwarding burden is maximum over the chain heads which causes their
early energy depletion and leads to shorten the network lifetime as well as stability
period. AUV is also incorporated that follows the predefined trajectory. It sup-
ports the MNs and GWs in data transmission towards the sink node. Sink node
stays at the top of water surface and responsible for data delivery at some terres-
trial station. Presence of AUV helps the other two type of nodes by shortening
their transmission distance. This results in energy conservation and shorter net-
work lifetime and stability period of the system. AUV passes by the sojourn stops
and stays there for some predefined sojourn time. It gathers the data from the
surrounding nodes that are finished with their data computation and processing.
42
5.5 Energy consumption model
S
C2
C1
D
S
C2
C1
D
Path 2
Path 1
Path 3
Case A Case B
3rd slice
2ndslice
1st slice
Sector
Ist Layer
2nd Laye
Figure 5.3: Cases of energy consumption model
Principal feature of our study is to model the energy consumption for man-
aging a packet communication event. For unique scenarios we group the energy
consumption as a vector. We employed the sensor nodes which are equipped with
transmission power control module [39–41]. Therefore the nodes are capable to
adjustable transmission powers. We suppose the uniform random deployment of
network. Hence, the energy consumption in data communication between slices is
constant for a sensor belonging to a particular slice. Consequently, the total energy
consumption by the sender nodes in data transmission towards the destination as
well as the cooperating nodes can be represented in the form of equations written
below.
Es1
S= 0.2,(5.5.1)
Es2
S= 0.4,(5.5.2)
Es3
S= 0.2 + 0.4 + 0.4 = 1,(5.5.3)
Equation[12] depicts that the maximum energy consumption is 1 for sending a
data packet inside its transmission range. In this section we will propose an energy
consumption model for three unique scenarios:
43
•−−→
EV for energy consumption of source node in sending data towards the three
slices.
•−−→
EV for energy consumption at different paths followed by data packets in
data transmission process.
•−−→
EV for energy consumption in data transmission and retransmission at dif-
ferent slices.
As an example consider the scenario depicted in Fig. 5.3 and assume a worst sce-
nario that all the three nodes have asked for a retransmission. For the above
mentioned specific scenario −−→
EV displays the energy being consumed by the sender
node in sending the data towards receivers residing in slice 1 (layer 2),2 and 3
respectively. We depicts −−→
EV with three vertices: from top to bottom 1st , 2nd and
3rd vertex depicts the energy consumption at layer 1 due to data packet forwarding
towards 1st , 2nd and 3rd slice respectively.
Energy Vector representation for case A is given as under:
−−−→
EV A
S=
0
0.42
0
+
0
0
1
+
0
0.42
0
,(5.5.4)
−−−→
EV A
S=
0
0.32
1
.(5.5.5)
Energy Vector representation for case B is represented as under:
−−−→
EV B
S=
0
0.42
0
+
0
0
12
+
0.22
0
0
,(5.5.6)
−−−→
EV B
S=
0.4
0.16
1
.(5.5.7)
44
Till then we have calculated the energy consumption at layer 1 in data forwarding
process towards different slices. Furthermore consider a case that we are intended
to calculate the total energy available at layer 1 (containing the sender node) after
the conduction of a packets transmission process.
For that purpose take into account the previously calculated energy vectors
and divide their entries with the total number of sensors belonging to 1stlayer.
Hence, for case A −−→
EVScan ne represented as:
−−−→
EV A
S=
0
0.32/E ×S1
1/E ×S1
,(5.5.8)
while, for case B it is calculated as follows:
−−−→
EV B
S=
0.4/E ×S1
0.16/E ×S1
1/E ×S1
.(5.5.9)
Whereas generalized −−−→
EV G
Sshows the energy available at the source node in sending
of data towards nodes residing in slice 1, 2 and 3 respectively. It is given as under:
−−−→
EV G
S=
E1
S/E ×S1
E2
S/E ×S1
E3
S/E ×S1
(5.5.10)
Till then we have modeled the energy consumption taking place at source node.
Besides that we are extending the energy consumption model for a different sce-
nario as during the data transmission inside a sector a packet propagates through
all the three slices. Therefore we are going to formulate an energy consumption
vector to represent the amount of energy being consumed at different paths fol-
lowed by the data packet in between the source and destination node.
45
It is to note here that previously we have formulated the energy consumption
model for both cases to enhance the readers understanding. However for now
onwards in order to keep the model simple we are going to consider the above
depicted case B only.
−−−→
EV T
Pdepicts three vertices, from top to bottom 1st, 2nd and 3rd vertex shows the
energy consumption in data forwarding through path 1,2 and 3 respectively. It is
depicted as under;
−−−→
EV T
P=
−−−→
EV B
P1
−−−→
EV B
P3
−−−→
EV B
P2
(5.5.11)
The passageway that is being followed by the data packets in different slices, shows
different energy consumption values. Hence total energy consumption at different
paths can be represented as:
−−−→
EV T
P=
0.42+ 0.42
12
0.22+ 0.82
=
0.32
1
0.68
,(5.5.12)
in the above equation 1st, 2nd , 3rd elements of vectors confirms the energy being
consumed in data propagation via first, second and third paths is 0.32, 1 and 0.68
respectively. Therefore, if it is required to follow a path with minimum energy
consumption then 1st path is best whereas, 3rd one is worst.
Till then we have elaborated i) the energy being consumed by source node in data
transmission towards the 1 to 3 slices and the three different possible paths for
data flow in between the source node and D. Now we are proceeding to depict the
energy being consumed by the nodes in different slices.
For case B −−−→
EV B
sis shown:
−−−→
EV B
s=
0.42+ 12+ 0.22
0.42
0
=
1.2
0.16
0
.(5.5.13)
46
Hence, from above equation we can deduce that the major portion of energy is
consumed at the slice containing source node and a minor portion is being con-
sumed at cooperating nodes but still we cannot ignore the reasonable portion of
energy that is being consumed at the other nodes except of source node. Hence, it
is also proved here that cooperative communication consumes a reasonable amount
of energy as compare to a non cooperative environment. It is also to note here that
no energy is being utilized at destination node because energy consumed due to
packet reception is very low as compare to the energy consumption due to packet
transmission. Thats why for the sake of our convenience we assumed it to be zero
in our scenario.
Therefore, total energy available in different slices can be depicted as:
−−−→
EV T
s=
2.2/E ×S1
0.16/E ×S2
0/E ×S3
,(5.5.14)
whereas, −−−→
EV G
scan be represented as follows,
−−−→
EV G
s=
E1/E ×S1
E2/E ×S2
E3/E ×S