Priority Enabled Distance-energy based Routing Algorithm

for UWSN

C. P. Gupta

, Mayank Bisht

and Arun Kumar

Department of Computer Engineering, Rajasthan Technical University, Kota, India

Department of Mathematics, Government College, Kota, India

Keywords: UWSN, Routing, Acoustic Signal, Routing Factor, Energy Scale Value, Priority Packets.

Abstract: Underwater Sensor Networks (UWSNs) are being deployed for range of applications like collection of

oceanic data for research, military surveillance, disaster prevention, underwater exploration etc.

Characteristics such as use of acoustic signal for communication, 3D deployment, and higher losses make

routing in UWSNs different from terrestrial sensor networks. In this paper, we present a location aware

routing algorithm based on routing factor (Rf); a function of distance and energy. In our proposed

algorithm, forwarding node is selected by sender amongst its neighbors depending on their distance from

destination node and residual energy. To consider energy with distance, Energy scale value (Es) is used as a

scaling range. Priority packets are also used for quick delivery of packets. Simulation results show improved

performance of our routing algorithm in terms of network lifetime and end to end delay.

1 INTRODUCTION

Underwater Sensor Networks (UWSNs) provide

huge potential for development & utilization of

underwater resources. Sensor nodes & Autonomous

Underwater Vehicles (AUVs) are envisioned to find

application in the exploration of underwater regions

for environmental monitoring, intrusion detection &

surveillance, mine detection, assisted navigation,

underwater exploration and seismic sensing (Hied.

et al, 2012). But these potential applications are

viable only if we have efficient underwater

communication system.

Characteristics of UWSNs differ from terrestrial

WSNs in terms of communication methods, network

deployment and protocols etc (Davis and Chang,

2012). Since radio waves suffer from high

attenuation in water, acoustic signals are used for

communication in UWSNs. This renders terrestrial

routing techniques unsuitable for UWSNs. UWSNs

also suffer from high delays, transmission losses and

node mobility due to water currents, which may

result in loss of connectivity and node failures

(Manjula et al., 2011). Routing protocols designed

for sensor networks are based on characteristics such

as type of signals used, available power &

bandwidth, delays, losses, node deployment (Zaihan,

2008). However, advancement in semiconductor

technology have overcome limitations of processing

speed, storage in UWSNs, still underwater

deployments occur over shorter periods (several

days), rather than months or years common in

terrestrial sensing. Efficient Routing techniques can

improve the lifetime of the network.

In this paper, we propose a routing algorithm that

considers both distance and energy of nodes for

making routing decisions in a 3-dimensional UWSN.

The proposal is a location based algorithm in which

all nodes are aware of their position in the network.

Routing decision is taken by the sender based on

Routing factor (Rf); a function of neighbour’s

distance to sink and its residual energy. High priority

packets are routed differently ensuring lower end to

end delay. Routing tables are used to reduce packet

transmissions among nodes and hence improve

performance. Our simulations show improvement in

lifetime & network throughput with satisfactory end

to end delays.

Rest of the paper is organized as follows. In

Section 2, we will review some existing routing

protocols for UWSNs. Section 3 describes our

proposed routing algorithm. Performance evaluation

of the proposed algorithm is presented in section 4.

Finally conclusions are drawn in Section 5.

133

P. Gupta C., Bisht M. and Kumar A..

Priority Enabled Distance-energy based Routing Algorithm for UWSN.

DOI: 10.5220/0004759001330138

In Proceedings of the 3rd International Conference on Sensor Networks (SENSORNETS-2014), pages 133-138

ISBN: 978-989-758-001-7

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

2 RELATED WORK

Vector Based Forwarding (VBF) (Xie et al., 2006) is

a location based routing protocol involving only a

fraction of nodes in routing. Packets are forwarded

along a virtual tunnel from source to sink. A self

adaptation algorithm for adjusting the forwarding

policy based on node density was also proposed.

The algorithm introduces desirableness factor in the

range of (0, 3) to measure the suitability of a node to

forward packets. Received packet is held by the

node for a time period related to its desirableness

factor, such that node with less desirableness factor

will forward the packet earlier. However, redundant

packet transmissions and packet delays cause energy

losses requiring alternate measures.

Focused Beam Routing (FBR) protocol (Jornet et

al., 2008) uses a distributed approach, in which route

is dynamically established as the data packet

traverses the network towards its final destination.

For finding all the nodes in a cone with ±θ/2

emanating from the source nodes towards the

destination nodes at the minimum distance, a

Ready_To_Send (RTS) signal with minimal energy

is transmitted. In case, no node responds through a

Clear_To_Send (CTS) like packet, the power level

and if required also value of θ is varied. The node

closer to final destination is selected as the relay

node for the next hop. However, performance of

algorithm is heavily dependent upon collision of

CTS packets at the source of RTS. End to end Delay

is also high in FBR.

Depth Based Routing (DBR) (Yan et al., 2008)

requires only local depth information against the full

location information required in VBF. DBR

assumes multiple sinks deployed at the surface

communicating with each other & Base Station

through radio links. Each packet in DBR contains

the depth information. On receiving a packet, node

forwards it only if it is closer to sink i.e. situated at

lower depth than sender node. Priority queue

mechanism is used to reduce the number of

forwarding nodes transmitting the same packet.

Each node receiving the packet compute packet

holding & scheduled sending time based on its depth

such that the node at lower depth transmit the packet

earlier than node at a larger depth. The algorithm

requires synchronization of clocks to ensure that

scheduled sending time is computed correctly by all

the nodes. Also, it requires specific deployment with

sink nodes floating on water surface.

An Energy Efficient Localization free Routing

Protocol named EEDBR proposed by Wahid et al.,

2012 also utilizes the depth of sensor nodes for

forwarding data packets along with the residual

energy of sensor nodes to improve the network

lifetime. Sender node enquires depth information

among its neighbours and according to their depths

create prioritized node list. On receiving packet,

each node holds the packet for some time on the

basis of its priority in the priority list. The EEDBR

results in improved network lifetime, energy

consumption and end-to-end delays and offers

comparable delivery ratio. However, the proposed

algorithm requires sorting for assigning priorities

which require more storage and computing power

within the sensor nodes. Also it requires specific

deployment with sink nodes floating on water

surface.

SBR-DLP (Sector-Based Routing with

Destination Location Prediction), proposed by

Chirdchoo et al., 2009 is also a location based

routing protocol for UWSN. SBR-DLP assumes sink

node to be mobile with pre-planned path and

schedule known to all other nodes in the network.

The whole range of node is divided into a number of

sectors. The sectors are prioritized based on angular

differences from the virtual vector SD from the

sender S to Destination D. Then according to sector

priority, the node closest to predicted location of the

mobile sink is selected as forwarder node. Latest

network information is acquired each time before

sending a packet using chk_ngb & chk_ngb_reply

packets. Limitations of this algorithm include large

delay between the packets due to

chk_ngb/chk_ngb_reply packets.

More routing techniques for UWSN are

discussed in (Wahid et al., 2010). Unlike location

unaware routing algorithms DBR and EEDBR, our

proposal does not require sinks to be floating on the

surface. Our algorithm works even if the sink is

mobile or at distant region of network deployment.

3 PROPOSED ALGORITHM

In our algorithm, routing decision depends on the

amount of residual energy of the neighbour node and

its distance from the destination node. Sender

decides the next forwarder node from its neighbours

and unicasts the packet to that node. Thus, our

algorithm attempts to route the packet through a

node which balances energy consumption in the

neighbouring nodes while maintaining acceptable

packet delay and delivery ratio. This avoids

selecting a certain node or group of nodes every

time to forward a packet. UWSN characteristics

such as 3-D network architecture, node mobility,

SENSORNETS2014-InternationalConferenceonSensorNetworks

134

acoustic channels, and limited power availability

have been taken into consideration in our proposal

Following assumptions have been made while

designing the algorithm:

 Each node knows its location. It is required as the

proposed algorithm location based routing

algorithm (Vijay and Choo, 2006).

 Sinks are mobile and are equipped with

navigational and propelling system (as like an

Autonomous Underwater Vehicle). The

trajectory of sinks is pre-planned and is known to

all the nodes in the network. Sinks are allowed to

deviate from the trajectory only within a range.

 All nodes other than sink node(s) have random

walk dynamic mobility pattern.

3.1 Packet Formats

Three types of packets namely; Hello, Ack and

Routing are used in our proposal. Hello Packet is

broadcasted by a node to enquire about its

neighbouring nodes in the network. Ack packet is

sent by a node in reply to Hello packet. After

receiving Ack, nodes update their neighbour table

with the information contained in this packet.

Routing Packet contains information about the

packet and data to be sent from source to sink node.

Priority_Bit sets priority with which packet is to be

sent by the sender. It is 1 for priority packet and 0

for normal packet. The packet formats are shown in

Fig 1.

Sender_Id

(10)

Position

(36)

Residual

Energy(16)

Broadcast

address (10)

(a) Hello Packet (72 bit)

Sender

Id (10)

Position

(36)

Residual

Energy(16)

Unicast

address (10)

(b) Ack Packet (72 bit)

Source

Id (10)

Forwarder

Id(10)

Sink

Id(10)

Packet_Sequence

Number(15)

Priority

Bit (1 )

Data

Figure 1: Packet Formats.

3.2 Routing Tables

Two tables are maintained by each node to minimize

exchange of control packet (Hello and Ack), speed

up packet transmission and reduce end to end

delays.

(a) Neighbour Table: Neighbour table holds

information of neighbours which is updated

whenever the node has a packet to forward and is

supposed to be stable for a predetermined duration

based on intensity of water currents. Higher duration

is set for networks deployed in still water. The

neighbour table has the following structure

<Neighbour_Id, Position, Residual_Energy>

(b) Sink Table: Sink table holds the information of

the sinks deployed in the network. It also stores

information related to previously taken path by a

packet from that node to each sink listed in table.

The sink table stores the following information:

<Sink_Id, Position, Fwd_Id_nomal, Lf

, Fwd_Id_priority,

Sink_Id and Position represents the position of sink

node; Fwd_Id_normal & Fwd_Id_priotity represents

the previous forwarder’s Id in normal and priority

modes; Lf

& Lf

are the time intervals called

lookup factor for the validating the suitability of

sending current packet through previous forwarder

node to sink node.

Algorithm: Distance Energy based Routing

Algorithm with Priority Handling.

At Each Node:

1. If nbr_table == empty() OR t

nbr_upd

is expired

Create nbr_table

i. Broadcast Hello packet

ii. Analyze Ack packets replied by nodes.

iii. Update nbr_table , t

nbr_upd

& d_thresh.

2. If a node has packets to send/forward

i. Create packet, set sink_id and priority.

ii. In sink_table against the sink_id and

packet_priority check Lf

iii. If ‘Lf ’ is not expired OR ‘Lf ’ ! = NULL

a. Schedule the packet for forwarding using

previous forwarder node in sink_table.

iv. If ‘ Lf ‘ is expired then

a. Calculate the Rf for each neighbour node

b. Select node with minimum Rf value as

forwarder node to send packet

c. Schedule the packet for forwarding.

d. Update sink_table

3. If a node receives a packet

i. If Hello packet then reply with a Ack packet

ii. Else if routing packet then

a. Extract source and sink information.

b. If node_Id == sink_Id then

i. Receive the packet

c. Else forward packet by following step 2.

3.3 Design of Algorithm

Design factors and elements of our algorithm are

discussed below:

(a) Routing Factor (Rf) and Energy Scale Value

(Es): Routing Factor (Rf) is computed by sender

PriorityEnabledDistance-energybasedRoutingAlgorithmforUWSN

135

node on the basis of distance between its neighbour

& sink and neighbour node’s residual energy such

that the most suitable node for forwarding the packet

has minimum Rf. Energy Scale Value (Es) is

scaling range for node’s residual energy to

commensurate it with distance for computing Rf.

Let, Distance between neighbour & destination

node = dist(n,d),

Energy Scale Value = Es,

Current residual energy level = E

res

Energy Difference or Initial Energy = E

diff

Then, Routing factor is given by:

Rf = dist(n,d) + Es * ( 1- E

res

diff

) (1)

This is the Rf for forwarding a normal packet. While

forwarding a priority packet, ‘Es’ is assumed to be

0. ‘I’ plays a major role in balancing the energies of

candidate forwarding nodes. It adds up an extra

value to Rf against node energy to make the routing

decision dependent on energy also.

Fig.2 illustrates routing decision based on Rf.

Distance d

is distance of destination node D to

neighbour node i and ed

is scaled value of energy of

neighbour nodes. Value of ed

is less for node having

high residual energy. From (1) we have Rf as the

sum of d

and ed

. The node with minimum Rf is

selected as next forwarding node by sender S.

Figure 2: Illustration of routing decision.

In above figure, neighbour node 2 has minimum

distance d

but less energy (as ed

is large) while

node 3 has more distance d

and more energy (as ed

is small). The overall Rf is less for node 2 hence,

node 3 is selected as forwarder. For a priority packet

as only distance is considered for computing Rf so

node 2 will be selected for forwarding the packet.

(b) Distance Threshold (d_thresh) and Packet Burst

Size (bs): Distance threshold is a function of time

used to cancel out the motion effect of nodes that

may move out of the range of sender before t

nbr_upd

expires. Whenever t

nbr_upd

is set, value of d_thresh is

set to minimum value. As the time progresses,

d_thresh increases. “d_thresh” is the maximum

motion of nodes after ‘t’ units of time. For finding

suitable forwarder then, node decreases its range by

d_thresh. Packets are generated by nodes in burst of

1 to 4 packets depending on size of information to

be sent. Packet can be generated by any node in the

network (except sink node).

decision is taken by the source or intermediate

sender node itself depending on type of packet, by

accessing the neighbour information in neighbour

table or previous forwarder information in sink

table. The packets can be forwarded as normal or

priority packet as decided by source. Priority packets

is an arrangement of sending packet with minimum

end to end delay by considering only distance

information for urgent information. Priority packets

are not targeted to balance energy and are given

priority among other packets at each node in the

network.

(d) Packet Acknowledgements: Acknowledgements

can be carried out in two ways. First way is to use

acknowledgements from sender node and forwarder

node in a hop by hop fashion. Secondly, we can use

packet acknowledgement from sink node to source

node on successful delivery of packet. Size of these

acknowledgements is very small so they can be

easily used in the network and can even be

piggybacked by other packets. However, due to

node motion, it is difficult to provide second type of

acknowledgements. Hence, we use hop by hop

acknowledgements to ensure successful delivery of

packets.

4 RESULTS AND ANALYSIS

We simulated our routing algorithm on a simulator

program created in C++ and use aqua3d animator to

visualize the simulations and working of our

algorithm (Tran, 2009). Simulations were performed

a large number of times and the results were

averaged from all results. Table 1 list the parameters

used in our simulations. For evaluation of our

proposed routing algorithm, we use following

performance metrics.

Lifetime: Network Lifetime is the time before first

node die in the network. We considered lifetime as

the time until a number of nodes die in the network.

End to End Delays: It is the time taken by a packet

to reach from source node to destination/sink node.

Packet Delivery Ratio: It is the ratio of number of

unique packets successfully delivered at the sink

SENSORNETS2014-InternationalConferenceonSensorNetworks

136

Table 1: Simulation Environment.

SIMULATION

SETTING

VALUE

Node Deployment Area 1000 x 1000 x 600 m

Deployment Type Random deployment

Node Speed

1 to 3 m/sec (random

direction)

Modem Type Acoustic Modem

Antenna Type Omni-directional

Transmission Range 300 metres

Data Rate 15000 bps

Speed of Sound 1500 m/sec

Size of Data Packet 4800 bit

Size of Control Packets

(Hello & Ack)

72 bit

Energy Scale Value 300

Packet Burst Size 1 to 4 packets

Neighbour Update Time

nbr u

)

4 sec

Lookup Factor (Lf

& Lf

) 2 sec

Number of Sink Nodes 6

Number of Total Nodes Variable ( 60 to 120)

4.1 Performance Evaluation

(a) Lifetime: We evaluated lifetime of the network

against percent of nodes dead in the network for

both the proposed routing scheme and SBR-DLP..

Comparison of overall lifetime in both the routing

schemes is shown in Fig 3. We observe an increase

in lifetime by a factor of 2 with respect to SBR-

DLP. This is because SBR-DLP always enquires

about neighbouring nodes before sending a packet

hence nodes die soon (Chirdchoo et al., 2009). Our

routing technique employs balanced energy

consumption and thus improves network lifetime.

The lifetime of SBR-DLP does not much deviate

even after increase in node density as the number of

transmissions to find neighbour nodes also

increases.

Figure 3: Comparison of Overall Lifetime with SBR-DLP

(b) Packet Delivery Ratio: Comparison of overall

PDR in both schemes is shown in Fig. 4 below.

Overall PDR decreases rapidly as the dead

nodes increase in the network. At lower network

densities, delivery ratio of SBR-DLP is much higher

than our proposed routing. However, at high

densities we observe a comparable overall PDR in

both the routing schemes. SBR-DLP performs well

in case of PDR than our routing because it utilizes

latest network information while performing but it

costs more energy usage and hence the lifetime of

the network (Chirdchoo et al., 2009).

Figure 4: Comparison of Overall PDR with SBR-DLP.

However, number of packets actually delivered

should also be considered before accounting for

higher delivery ratio in SBR-DLP. Because more the

number of packets send, higher are the chances of

packet loss, hence lesser delivery ratio. Fig. 5 shows

packets delivered in both routing schemes. Number

of packets delivered increases with increase in

network density. In our routing scheme number of

packets delivered is much more compared to that in

SBR-DLP. In SBR-DLP, energy drains out in

successive transmissions in finding network

information before sending each packet. Increase in

number of packets generated and delivered

decreases the delivery ratio in our proposed routing

algorithm.

Figure 5: Comparison of Number of Packets delivered

with SBR-DLP.

routing schemes is shown in Fig 6. We observe

comparable delays in both the routing schemes. At

PriorityEnabledDistance-energybasedRoutingAlgorithmforUWSN

137

adequate node densities our algorithm performs

better than SBR. However, at low densities, due to

inadequate routing options we observe some

increase in end to end delay. Also, delays are

dependent on node motion and neighbor density

which can be highly unpredictable at times.

Figure 6: Comparison of Overall Average E2E Delay with

SBR-DLP.

5 CONCLUSIONS AND FUTURE

WORK

Energy efficiency is one major issue in UWSNs. In

this paper we proposed distance energy based

routing algorithm which improves the lifetime of

underwater networks by utilizing location and

residual energy information to route a packet.

Simulation results shows that the proposed routing

algorithm improves the network lifetime with

satisfactory packet delivery ratio and end to end

delays. Also it has the priority concerns for a packet

which allow the packet to be forwarded with the

shortest path possible with high priority and minimal

waiting time. The simulation results are analysed on

various performance metrics and the results were

satisfactory.

Our routing algorithm needs to be developed

further so that it complies with optimality

constraints on Energy Scale and mobility. Complex

routing scenarios like void prevention, looping of

data packets, packet collisions also need to be

addressed. Other improvements include minimizing

overheads, increase channel utilization, self

configuring nodes, incorporating the localization

algorithm as a part of routing algorithm. Developing

this algorithm as a part of open source software like

NS2 will make it susceptible with networking

standards.

REFERENCES

Chirdchoo, N.; Wee-Seng Soh; Kee Chaing Chua, 2009,

“Sector-Based Routing with Destination Location

Prediction for Underwater Mobile Networks”, IEEE

international conference on Advanced Information

Networking and Applications, WAINA’09, pp. 1148-

1153, 2009.

Davis, A.; Hwa Chang; 2012, "Underwater wireless sensor

networks," Oceans, 2012, vol., no., pp.1-5, 14-19,

Heidemann, John, et al, 2012 "Underwater sensor

networks: applications, advances and challenges."

Philosophical Transactions of the Royal Society A:

Mathematical, Physical and Engineering Sciences

370.1958: 158-175.

Jornet, Josep Miquel, Milica Stojanovic, and Michele

Zorzi; 2008, "Focused beam routing protocol for

underwater acoustic networks.", Proceedings of the

third ACM international workshop on Underwater

Networks. ACM, 2008.

Manjula.R.B, Sunilkumar S. Manvi, 2011, “Issues in

Underwater Acoustic Sensor Networks ”, International

Journal of Computer and Electrical Engineering.

Tran, Matthew T., 2009, “Aqua3D: Three-Dimensional

Animator for Underwater Sensor Networks”,

University of Connecticut, 2009.

Vijay Chandrasekhar, Yoo Sang Choo, 2006,

“Localization in Underwater Sensor Networks –

Survey and Challenges” In proc. of WUWNet’06, Los

Angeles, California, USA, September 2006.

Wahid, A; Dongkyun, Kim; 2010, “Analyzing Routing

Protocols for Underwater Wireless Sensor Networks”,

International Journal of Communication Networks and

Information Security (IJCNIS) , December 2010.

Wahid, A; Dongkyun, Kim; 2012, “An Energy Efficient

Localization-Free Routing Protocol for Underwater

Wireless Sensor Networks,” International Journal of

Distributed Sensor Networks, vol. 2012,

Xie, Peng, Jun-Hong Cui, and Li Lao. 2006, "VBF:

vector-based forwarding protocol for underwater

sensor networks." Networking 2006. Networking

Technologies, Services, and Protocols; Performance of

Computer and Communication Networks; Mobile and

Wireless Communications Systems (2006): 1216-

1221.

Yan, Hai, Zhijie Shi, and Jun-Hong Cui. 2008, "DBR:

depth-based routing for underwater sensor networks."

NETWORKING 2008 Ad Hoc and Sensor Networks,

Wireless Networks, Next Generation Internet (2008):

72-86.

Zaihan Jiang, 2008, “Underwater Acoustic Networks –

Issues and Solutions”, IJICS 2008.

SENSORNETS2014-InternationalConferenceonSensorNetworks

138