Introducing a novel fault tolerant routing protocol in wireless sensor networks using fuzzy logic

International Journal of Computer Science & Information Technology (IJCSIT) Vol 5, No 5, October 2013

INTRODUCING A NOVEL FAULT TOLERANT
ROUTING PROTOCOL IN WIRELESS SENSOR
NETWORKS USING FUZZY LOGIC
Mostafa Vakili Fard 1 , Sayyed Majid Mazinani2 , Sayyed Ahmad Hoseini3
1

Department of Computer Engineering, Science and Research IAU , Neyshabur, Iran
2
Department of Computer Engineering, Imam Reza University , Mashhad , Iran
3
Department of Computer Engineering, , payam Noor University , Mashhad , Iran

ABSTRACT
Although routing is one of the key network issues, the challenges facing WSN makes the subject even more
demanding. In this study, we propose a novel fault tolerant routing protocol in wireless sensor networks
using fuzzy logic. We approach the problem by breaking it down into two parts. The first sub problem is
designing a routing module and the second one retrofitting the designed module in a way that answering to
these two sub problems leads to the final solution of the main problem. To deal with the first stage, by using
fuzzy logic, we design a routing module; then for the second stage, the module enters fault tolerant routing
in order that it be tolerant against routing faults. In this study, the term "network failure" is figuratively
equal to "node failure".

KEYWORDS
wireless sensor networks, routing, fault tolerance, fuzzy logic.

1. INTRODUCTION
Routing in wireless sensor networks embraces a different concept comparing to other wireless
communication networks due to a handful of fundamental characteristics; therefore, the
researchers have faced new challenges. The very first challenge in wireless sensor networks is
incapability of building a unique global address for each node. Consequently, the protocols, used
in those wireless networks functioning under IP protocol, will be considered useless in such
networks. The second challenge is that nearly all applications under WSN require transferring the
sensed data from the surroundings to a certain unit, called the base station. As for the third
challenge, the data traffic generated in these networks has a considerable redundancy. One event
can be detected by different nodes and they accordingly send the same data. Such redundancies
should be considered in order to use energy consumption in nodes and bandwidth more logically.
The fourth challenges concern the limitations in data-saving energy, power computation, and
node total capacity[1].

2. R EVIEW OF LITERATURE
Fuzzy logic is used to recognize the values which are not one hundred percent true or false, and
which are facing against classical logic that has been utilized in control systems thus, they are
widely used in various systems, such as WSNs to name a few. Algorithm of [2] introduces a
routing protocol which by using fuzzy logic obtains the node’s residual energy, local density
DOI : 10.5121/ijcsit.2013.5513

171


inside its sensed region, and time as input parameters of fuzzy module. This leads to a fair
distribution in order to make a cluster. This protocol benefits from two fuzzy modules. Algorithm
of [3] presents a routing protocol which selects the proper cluster head based on fuzzy logic.
Selecting the proper cluster head has an impact on energy saving and extending the network
lifetime. By using three parameters of energy, density and centricity this algorithm makes better
decisions of fuzzy modules. In algorithm of [4], the head clusters are selected by the base station.
Fuzzy controller is our best routing component. Its input parameters are the head cluster’s center
of gravity , node’s remaining energy, and the network traffic. Algorithm of [5] presents a fault
tolerant routing mechanism which uses a multi-path technique. Such algorithm uses separate
paths in order to deliver the sensed data to the base station. Doing so, it selects a supporting head
cluster for each head cluster. Algorithm of [6] presents a routing protocol with energy efficiency
which uses the fixed delay and selects the next hop proactively utilizing both paths.
In this study, we propose a novel fault tolerant routing protocol in wireless sensor networks using
fuzzy logic. Benefiting fault tolerant routing protocol paves the way of carrying on the routing
operation in the presence of faults. Here, the term "network failure" is figuratively equal to "node
failure".

3. PROBLEM DEFINITION AND SOLVING
We are breaking the problem down into two sub problems of designing a routing module based
on fuzzy logic and of retrofitting it. We are hypothesizing flat network architecture and querybased network operation, along with a proactive routing and routing will happen at the time of
generating the sensed data.

3.1. Designing Routing Based On Fuzzy Logic
The first hop to solve the early sub problem, the fuzzy module is designed based on the Mamdani
fuzzy interference system. In Mamdani system, input and output are expressed in fuzzy. Figure 1
illustrates parts of a fuzzy system.

Fig 1. Fuzzy inference system components

Four parameters Residual energy of node, path traffic, numbers of hops to the base station, and
packet priorities are input parameters of the fuzzy module. These parameters enter the fuzzy
module through neighbour management unit (NMU).
172


Figure 2 illustrates the function membership related to the node's residual energy. Table 1
illustrates the information concerning the other membership functions of fuzzy module.

Fig 2. Residual energy membership function

According to figure 1, the fuzzy rules are used as one of fuzzy system parts. The following is the
general idea of a fuzzy rule:
IF Variable Is Set Then Action
Since we have used the fuzzy module in routing designing, to receive satisfying output, we utilize
center-of-gravity function in order to dismiss the output from its fuzzy mode. When a node
contains sensed data, the NMU provides the requiring information for the fuzzy routing module.
The output parameter is considered to be as the node goodness. Thus, the fuzzy inference engine
calculates the node goodness value and singles out the most suitable node.
Table 1. Fuzzy module membership function

Path Traffic(%)
Number Of Hops
Packet Priority
Node Goodness

Low
40< <0

0 < <3
0<<1
0<<0.5

Medium

High

30 < < 80
2<<4

70<<100
3<<5
1.5 < <2
0.6<<1

0.5 < < 1.5

0.25<<0.75

3.2. The Fault Tolerating Step
As mentioned earlier, the fault tolerating aims to let the routing operation happen in the presence
of faults and, finally, the packets be delivered to the destination. On the other hand, redundancy is
the very basis of retrofitting. So, we devise an alternate path besides the main path for each node's
sensed data down to the destination. In this stage, the output taken from fuzzy routing designing
is considered to be a black box. Of all the neighbour nodes, the one with maximum membership,
selected as the next hop, places in the first best path and the node with the second ranking does in
the second best path. The first best path is naming the main path and the second best one the
alternate path. The NMU collects data relating to each node's membership in main and alternate
paths. For a generating node, it's only possible to be a member of either the main or alternate path.
Figure 3 illustrates our proposed routing mechanism. The yellow node is the node ready to send
data; the main path is in green and the alternate path is in red arrows.

173


Fig 3. Proposed Routing Mechanism

4. SIMULATING OPERATION
As mentioned in fuzzy routing input, the priority of the packets is one of the key input parameters
to the fuzzy module.
The packets are prioritized into three levels. Level Zero, the interested message distributing from
the base station to the network; Level One, the packets distributing among the nodes and
management neighbour units; Level Two, the sensed data by nodes. Levels with higher numbers
establish superior priorities.
We assign three variants to the suggested protocol, and name Level 0, Level 1, and Level 3, as
FTRF 0, FTRF 1, and FTRF 2 accordingly. In a way that the variant FTRF 0 assigns alternate
paths for Levels 0, 1, and 2.
Having done the simulation in Castalia environment, we have compared three parameters of the
network lifetime, consumed energy, and delivery ratio packets to the base station resulted from
the mentioned variants. Figures 4, 5, and 6 illustrate the results taken from three variants. After
studying the results, we pick the variant with the best result. Figure 4 illustrates scales of the
sensed packets comparing to the residual-energy average. Gradually, with increasing of numbers
of the sensed packets, and due to only one sort of redundant data, the variant FTRF 2 minimizes
its consumption to a moderate level, in comparison to other two variants generating more
redundant packets. Count ratio of sensed packets with percentage of critical nodes are used in
order to calculate the network lifetime.
Figure 5 illustrates that with increasing of numbers of the sensed packets, the critical nodes in
variant FTRF0 escalate, but the variant FRTF2 contains less critical nodes.
Figures 6 illustrates the comparison of parameters of the packet delivering in these three variants,
where by increasing of sensed data generating nodes, packet delivering to the base station reduces
-- a natural circumstance due to gradual consumption of nodes' energy. However, among the
variants, FTRF2 benefits from a better delivery ratio than the others.

174


Fig 4. Residual energy ratio with sensed packets

Fig 5. Critical nodes ratio with sensed packets

175


Fig 6. Delivery ratio with sensed data generating nodes

5 CONCLUSION
In this study, we propose a novel fault tolerant routing protocol in wireless sensor networks using
fuzzy logic. After designing and routing based on fuzzy logic, we utilize multi-path redundancy
and routing techniques in order to increase the suggested fault tolerant routing. Since, in this
study, the term "network failure" is figuratively equal to "node failure", the fault tolerant routing,
even facing node failures, must be capable of routing the packets to their final destination. The
variant FTRF2 features longer lifetime and delivers more packets to their destination among the
proposed redundancy of other variants.
In this study, there is no citation of recovering mechanism for routing failure. To ensure accuracy,
moreover, the membership functions of fuzzy module input can increase in order to provide
service for traffic requiring special service qualities.

REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]

K. Akkaya and M. Younes , “ A survey on routing protocols for wireless sensor networks ” , Ad Hoc
Networks , vol. 3 , no. 3 , May 2005 , pp. 325 – 349
Y.M Tashtoush and M.A Okour, Fuzzy Self-Clustering for WSN, IEEE/IFIP International Conference
on Embedded and Ubiquitous Computing 2008.
J.Kim, S.park, Y.Han and T.Chung, “CHEF:cluster head election mechanism using fuzzy logic in
WSN” , in proceeding of the ICACT , 2008,pp.654-65
Joupe Anno ,Leonardo Barolli , “ A cluster head selection method for WSN base on fuzzy logic” ,
IEEE 2007 , pp. 337 – 348.
D. Ganesan, R. Govindan, S. Shenker, and D. Estrin, “Highly-resilient, energy-efficient multipath
routing in wireless sensor networks,” vol. 5, pp. 11–25, Oct. 2001.
Yuan, L.F.; Cheng, W.Q.; Du, X. An energy-efficient real-time routing protocol for sensor networks.
Comp. Commun. 2007, 30, 2274–2283.

176

Introducing a novel fault tolerant routing protocol in wireless sensor networks using fuzzy logic

More Related Content

What's hot (19)

Viewers also liked (19)

Similar to Introducing a novel fault tolerant routing protocol in wireless sensor networks using fuzzy logic (20)

Recently uploaded (20)

Introducing a novel fault tolerant routing protocol in wireless sensor networks using fuzzy logic