011001010 a

Dr. Sohan Garg et al. / International Journal of Engineering Research and Applications
(IJERA) www.ijera.com Vol. 1, Issue 1, pp.001-010

A Fuzzy Logic Based Network Dependent Routing Algorithm for
Ad hoc Wireless Networks
Dr. Sohan Garg1, Payal Kansal2, Viksit Kumar Sharma3
( 1Department of Computer Applications, Raj Kumar Goel Institute of Technology,
Delhi-Meerut Road, GZB (U.P.)
2
Department of IT, Lord Krishna College of Engineering Ghaziabad
3
Rajeev Academy for Technology and Management, Mathura (UP) )

Abstract: The development of Mobile Ad Hoc G (v, e), with a set of vertices v and a set of edges e.
network advocates self-organized wireless
Each vertex of the set v represents a network node and
interconnection of communication devices that
would either extend or operate in concert with the each edge of the set e represents a wireless link. The
wired networking infrastructure or, possibly, evolve total number of nodes is n = |v|. This key feature of
to autonomous networks. Unlike traditional wireless
networks, ad hoc networks do not rely on any fixed these networks enable them to be employed in places
infrastructure. Instead, hosts rely on each other to where an infrastructure is not available, such as in
keep the network connected. One main challenge in
disaster and on battle grounds, or we can say that the
design of these networks is their vulnerability to
security attacks. Despite the existence of well-known dynamic nature of these networks and the scarcity of
security mechanisms, additional vulnerabilities and
bandwidth in the wireless medium, along with the
features pertinent to this new networking paradigm
might render such traditional solutions limited power in mobile devices (such as PDA’s and
inapplicable. In particular, the absence of a central laptops) make routing in these networks a challenging
authorization facility in an open and distributed
communication environment is a major challenge, task. The topology of an ad hoc network changes due
especially due to the need for cooperative network to the movement of mobile hosts, which may lead to
operation. In MANET, any node may compromise
sudden packet losses and delays. In the mobile ad hoc
the routing protocol functionality by disrupting the
route discovery process. In this paper, we network due to the high mobility, low signal power and
understand the various routing problems related to
limited bandwidth the wireless links are frequently
bandwidth, signal power, mobility and delay. In this
paper we are proposing a new routing algorithm broken and new links are frequently established. Such
that is totally network dependent and will remove dynamic network topology presents a significant
the all routing problems
challenge for the network routing algorithms. Several
Keywords: Delay, Mobility, Mobile Ad hoc routing algorithms, such as shortest path routing

network, topology algorithm like DBF were proposed for ad hoc wireless
networks. But, these algorithms suffer from very slow
1. Introduction
convergence (the “routing to infinity” problem).
Ad hoc networks are made up of a number of nodes, Besides, DBF-like algorithms incur large update
which are capable of performing routing without using message penalties. Protocols that attempted to cure
a dedicated centralized controller as a base station. Ad some of the shortcomings of DBF, such as Destination-
hoc networks can be represented as a connected graph Sequences Distance Vector routing were proposed.

www.ijera.com 01

Dr. Sohan Garg et al. / International Journal of Engineering Research and Applications (IJERA)
www.ijera.com Vol. 1, Issue 1, pp.001-010

However, synchronization and extra processing the algorithm is simple as well as having low
overhead are common in these protocols. Other communication overhead and storage requirements.

protocols that rely on the information from the The algorithm has been simulated in MATLAB 7.2 in

predecessor of the shortest path solve the slow which we have uses the Fuzzy Logic tool kit.

convergence problem of DBF. However, the
processing requirements of these protocols may be The Fuzzy Logic is an innovative approach to help

quite high, because of the way they process the update control non-repeating or unpredictable systems control

message. Realizations of the path findings algorithms, accuracy. It uses a list of rules rather than complicated

like the wireless routing protocol, are able to eliminate mathematical expression. Fuzzy Logic was introduced

the “counting-to-infinity” problem and reduce the by L.A. Zadeh in 1965. Fuzzy Logic is also known as

occurrence of temporary loops, often with less control fuzzy rule based system and this is a non linear

traffic than traditional distance vector schemes. The mapping technique of input data into output. Fuzzy

main disadvantage of WRP is the fact that routing Logic system is composed of five functional blocks.

nodes constantly maintain full routing information in 2. The Proposed Routing Algorithm
each network node, which was obtained as relatively
high cost in wireless resources. In the Temporary Here Fuzzy Logic has been used for routing and
Ordered Routing Algorithm the resulting route replies management of an ad hoc wireless network. The
are also flooded, in a controlled manner, to distribute proposed fuzzy logic based routing algorithm takes into
routes in the form of directed acyclic graphs (DAGs) account of three input variables, signal power and
rooted at the destination. In contrast, protocols such as mobility and delay. The absolute value of each of these
Dynamic Source Routing and ad hoc on demand parameters can take a large range at different points on
distance vector unicast the route reply back to the the network. We have considered the normalized
querying source, along a path specified by a sequence values for each parameter. Now, ‘crisp’ normalized
of node addresses accumulated during the route query values are being converted into fuzzy variables. For
phase. In the case of DSR, the node addresses are this, three fuzzy sets have been defined for each
accumulated in the query packet and are returned to the variable. The sets, low (from 0 to 0.4), medium (from
source, to be used for source routing. AODV, on the 0.2 to 0.8) and high (from 0.6 to 1.0) have been used
other hand, distributes the discovered route in the form for the input variable signal power (figure 1) and the
of next-hop information stored at each node in the sets, poor (from 0 to 0.4), average (from 0.2 to 0.8) and
route. These algorithms do not satisfy the requirements excellent (from 0.6 to 1.0) have been used for input
of an ad hoc wireless network completely and despite variable delay (figure 2) and the sets, low (from 0 to
their shortcomings, these works lay the foundation for 0.4), medium (from 0.2 to 0.8) and high (from 0.6 to
the development of our protocol. In this paper we 1.0) have been used for the input variable mobility
propose a new routing protocol for ad hoc wireless (figure 3). The normalized value of each parameter is
networks, which address some of the problems with the mapped into the fine sets. Each value will have some
existing approaches. In spite of all these improvements, grade of membership function for each set. The

www.ijera.com 02


memberships that have been defined for each of the protocol is distributed in nature without relying on
fuzzy set for any particular input variable are triangular centralized authorities.
in shape. Next the rules of inference have been written. 4. Entering/Departing nodes: This protocol is able to
Initially total 27 rules were devised. The crisp value of quickly adapt to entering or departing nodes in the
input variable was given and a defuzzified crisp value network, without having to restructure the entire
for selected variable was calculated from the derived network.
algorithm. An output linguistic variable is used to 5. On Demand operation: Since a uniform traffic
represent the route. Proposed optimal routes are based distribution cannot be assumed within the network, the
upon the fuzzy rules for different ranges of the metric routing algorithm must adapt to the traffic pattern on a
availability. The routes (figure 4) are defined as below demand or need basis, thereby utilizing power and
optimal (from 0 to 0.4), suboptimal (from 0.2 to 0.8) bandwidth resources more efficiently.
and optimal (from 0.6 to 1.0) between two mobile Our problem is to find the optimal and suitable route
hosts. The below optimal indicates not optimal path, from source to the destination based on mobility, signal
the sub optimal indicates good path and the optimal power and delay. The system is based on the fuzzy
path indicates the best path. The proposed routing inference system. The major components of the system
algorithm can apply to different routing metrics. These consist of the knowledge base, decision making,
routes have to satisfy the mobility, signal power and fuzzification and defuzzification. Now we will write
delay requirements of the network. The grade of the fuzzy rules based on the mobility, signal power and
membership function can be any where between 0 and delay and try to find out the optimal path or route.
1 for each fuzzy set. The defuzzified crisp value for
selected variable was calculated from the derived
algorithm. The proposed Fuzzy Logic & bandwidth
based routing algorithm for ad-hoc wireless network is
classified as “A Fuzzy Logic Based Network
Dependent Routing Algorithm for Ad hoc Wireless
Networks”. This algorithm contains the following
characteristics:
1. Freedom from loops: In the proposed algorithm the
paths derived from the routing tables of all nodes have
no loops.
2. Fast Route Convergence: The new routing protocol
provides a new and stable route as soon as possible
Figure 1. Input Variable ‘Signal power’
after a topology change.
3. Distributed Implementation: As we know those ad
hoc networks are autonomous and self-organizing, this

www.ijera.com 03


The Fuzzy Inference rules for the proposed routing
algorithm are following:
1. If (signal power is low) and (mobility is low) and
(delay is poor) then the route will be below optimal.
(delay is average) then the route will be below optimal.
(delay is excellent) then the route will be sub optimal.
4. If (signal power is low) and (mobility is medium)
and (delay is poor) then the route will be below
Figure 2. Input Variable ‘Delay’
optimal.
and (delay is average) then the route will be sub-
optimal.
and (delay is excellent) then the route will be sub
optimal.
7. If (signal power is low) and (mobility is high) and
(delay is poor) then the route will be below optimal.
8. If (signal power is low) and (mobility is high) and
(delay is average) then the route will be below optimal.
Figure 3. Input Variable ‘Mobility’ 9. If (signal power is low) and (mobility is high) and
(delay is excellent) then the route will be sub-optimal.
10. If (signal power is medium) and (mobility is low)
and (delay is poor) then the route will be sub- optimal.
and (delay is average) then the route will be optimal.
and (delay is excellent) then the route will be optimal.
13. If (signal power is medium) and (mobility is
medium) and (delay is poor) then the route will be
below optimal.
medium) and (delay is average) then the route will be
Figure 4. Output Variable ‘Route’
sub optimal.

www.ijera.com 04


medium) and (delay is excellent) then the route will be
optimal.
16. If (signal power is medium) and (mobility is high)
and (delay is poor) then the route will be below
optimal.
and (delay is average) then the route will be sub
optimal.
and (delay is excellent) then the route will be optimal.
19. If (signal power is high) and (mobility is low) and
(delay is poor) then the route will be sub-optimal.
20. If (signal power is high) and (mobility is low) and Figure 5. ‘Route’ O/P w.r.t. ‘Mobility’ and

(delay is average ) then the route will be optimal. ‘Signal Power’

21. If (signal power is high) and (mobility is low) and
(delay is excellent) then the route will be optimal. In figure 5 signal power and mobility are the

22. If (signal power is high) and (mobility is medium) fuzzy input variable for the proposed routing

and (delay is poor) then the route will be sub-optimal. algorithm which lies on the horizontal axes and

23. If (signal power is high) and (mobility is medium) route is the output variable which has been shown

and (delay is average) then the route will be sub- on the vertical axis. It is very clear from the figure

optimal. that at constant signal power and low mobility the

24. If (signal power is high) and (mobility is medium) Route is below optimal, but for any value of the

and (delay is excellent) then the route will be optimal. signal power if we increase the mobility then the

25. If (signal power is high) and (mobility is high) and route also increases and becomes sub-optimal.

(delay is poor) then the route will be below - optimal. Finally at high mobility the route will be below

26. If (signal power is high) and (mobility is high) and optimal for constant signal power.

(delay is average) then the route will be sub-optimal.
27. If (signal power is high) and (mobility is high) and
(delay is excellent) then the route will be optimal.

The routes based upon the above rules have been
shown with the help of the following graphs:

www.ijera.com 05


poor delay. But as the delay becomes high, route
Figure 6 ‘Route’ O/P w.r.t. ‘Signal Power’ and increases up to sub optimality and optimality when the
‘Delay’ mobility is medium and high respectively.

In figure 6 the inputs of the algorithm (signal power
and delay) are on the horizontal axes and the output
(route) is on the vertical axis. According to the figure It
is clear that at constant delay and low signal power the
route is below optimal but if the signal power is
increased up to medium the route is sub optimal for
poor delay. In the same way at high signal power and
poor delay the route becomes below optimal. If delay is
increased up to average at high signal power then the
route is suboptimal but route becomes optimal at
excellent delay and at high signal power.

Figure 8. ‘Route’ O/P w.r.t. Medium ‘Signal Power’,
Low ‘Mobility’ and Excellent ‘Delay’
The figure 8 illustrates that when signal power is
medium (0.504), mobility is low (0.0812) and delay is
excellent (0.885) then in this condition the route is
optimal (0.861). So this algorithm works well when
mobility is low and signal power is medium.
Figure 7. ‘Route’ O/P w.r.t. ‘Mobility’ and ‘Delay’

In figure 7 delay and mobility are the input variables
and route is the output variable. The 3 D decision
surface illustrate that route is below optimal for poor
delay and low mobility. As the mobility is increased up
to medium the route will be sub optimal. Again the
similar process occurs and route is below optimal for

www.ijera.com 06


Figure 9. ‘Route’ O/P w.r.t. Medium ‘Signal
Power’, Medium ‘Mobility’ and Excellent ‘Delay’ Figure 10. ‘Route’ O/P w.r.t. Medium ‘ Signal
In figure 9 we have seen that route is optimal at low Power’, Low ‘Mobility’ and Average ‘Delay’
mobility, medium signal power and at excellent delay This is very clear from the figure 10 that in this
but against it when we increase the mobility towards protocol the value of the signal power (0.47) and
medium (0.517) at medium signal power (0.504) and mobility (0.115) is same as in figure 6.8 but the
excellent delay (0.919) the route is again optimal. So difference is that when we will decrease the delay up to
this algorithm works well at medium mobility also. average (0.509) then the route is still optimal, but as we
will increase the mobility then the route will be sub
optimal.
3. Simulation Results
In decision surface (figure 5) the input variables signal
power and mobility lie on horizontal axes and the

www.ijera.com 07


output variable route lies on vertical axis. It shows that again but as we increase the mobility then route
at constant signal power the route is below optimal for becomes suboptimal.
low mobility. But any signal power if we increase the 4. Conclusions
mobility the route also starts increasing and becomes Due to the unprecedented growth in the scale and
sub optimal. Finally when the mobility becomes high, diversity of mobile computing devices, new
the route will be below optimal for constant signal horizons for wireless connectivity has come into
power. Now as we increase the signal power the route view. In this chapter, we have shown the
starts increasing towards optimality and becomes importance of an ad hoc routing protocol and some
optimal for low and medium mobility. In figure 6 the of the previous works. Following that we have
inputs of the algorithm are on the horizontal axes and proposed our new routing protocol based on
the output of the algorithm is on the vertical axis. Here, mobility, signal power and delay, where the
with the help of 3D surface we observe that at constant segmentation of nodes will substantially reduce the
delay and low signal power the route is below optimal overhead of the entire network and speed up the
but if the signal power is increased up to medium the routing process. After fully describing its functions
route is sub optimal for poor delay. Similarly at high and mechanism, we have suggested various
signal power and poor delay the route becomes below optimizations to the protocol and utilized the
optimal. It can also be observed that with average delay concept of stability index. Finally, we have done
and high signal power the route is suboptimal and if we limited trials to show that our protocol is functional
increase delay up to excellent the route is optimal at and effective; we do see the need in further
high signal power. In figure 7 the input variables experimentation in order to accurately access the
mobility and delay lie at horizontal axes and the route practical effectiveness of our protocol in a medium
at vertical axis. Here, we see that for poor delay and to large size network.
low mobility the route is below optimal. As the
References:
mobility increases and arrive up to medium the route
[1] A. James Freebersyser, B. Leiner, “A DoD
will be sub optimal. Again the similar phenomenon
perspective on mobile ad hoc networks”, in:
occurs and route is below optimal for poor delay. But
Charles E. Perkin (Ed.), Ad hoc Networking,
as the delay becomes high, route increases up to sub
Addison Wesley Reading, May 2001, pp.29-51.
optimality and optimality when mobility is medium
[2] J. Macker and S. Corson, “Mobile Ad hoc
and high respectively. In figure 8 we see that at
Networks (MANET)”, IETF WG
medium signal power (0.504), low mobility (0.0812)
Charter.,http://www.ietf.org/html.charters/manetC
and excellent delay(0.885) the route is optimal (0.86)
arter.html,1997.
and as we increase the only mobility towards
[3] IEEE Computer Society LAN MAN Standards
medium(0.517) then the route is again optimal (0.865)
Committee,”Wireless LAN medium access control
which has been shown in the figure 9. In figure 10,
(MAC) and physical layer (PHY) specifications”,
with medium signal power (0.47), low mobility (0.115)
IEEE Std. 802.11, 1997, pp. 11-97.
and average delay (0.509) the route becomes optimal

www.ijera.com 08


[4] M.S. Corson, J.P. Maker, J.H. Ernicione “Internet [13] J.J. Garcia Aceves,”Distributed routing algorithm
based mobile ad hoc networking”, IEEE Internet using internidal coordination”, Proc. IEEE
Computing, 3 (4), 1999, pp. 63-70. INFOCOM’, New Orleans, LA, 1998, pp.85-96.
[5] E. M. Royer and C. K. Toh, “A review of current [14] E.M. Gafni and D. P. Bertsekas,” Distributed
routing protocols for ad hoc mobile wireless algorithm for generating loop-free routes in networks
networks”, IEEE Personal Communications Magazine, with frequently changing topology”,
April 1999, pp. 46-55. IEEE Transactions on Communications, Vol. Com-29,
[6]I. Chlamtac and A. Lerner, “Link allocation in mobile no. 1, Jan. 1981
radio networks with noisy channel”, In IEEE , pp. 11-18.
INFOCOM, Bar Harbour. [15] M. S. Corson and A. Ephremides, “A distributed
www.openu.ac.il/Personal_sites/anat-lerner.html, FL, routing algorithm for mobile wireless networks”, ACM
April 1986. Journal for Wireless Networks, Feb. 1995, pp. 61-81.
[7] I. Chlamtac and A. Lerner, “Fair algorithms for [16] C.E. Perkins and P. Bhagwat, “Highly dynamic
maximal link activiation in multi-hop radio networks”, destination sequenced distance –vector routing (DSDV)
IEEE Transactions on Communications COM-3, Issue- for mobile computers,” ACM SIGCOMM, Vol.24, no.4,
7, Vol. 35, 1987, pp. 739-746. Oct.1994, pp. 234-244.
[8] A. Adya, P. Bahl, J. Padhye, A. Wolman, and L. Zhou. [17] S. Murthy and J.J. Garcia-Luna-Aceves, “An efficient
“A Multi –radio unification protocol for IEEE routing protocol for wireless networks,” ACM Mobile
802.11 wireless networks”. In IEEE International Networks and App. J., Special issue on Routing in
conference on broadband networks, (BroadNets), 2004. Mobile communication Networks, Oct. 1996, pp. 183-
[9] C. Siva Ram Murthy and B.S. Manoj, “Ad hoc 187.
Wireless Networks Architecture and Protocols”, [18] J.M. McQuillan, I. Richer, E. Rosen, “The new routing
Prentice Hall, 2004. algorithm for ARPANET”, IEEE Transaction of
[10] M.S. Corson, S. Batsel and J. Macker, “Architecture Communication Vol. 28, Issue May 1980, pp. 711-719.
consideration for mobile mesh networking”, [19] D.B. Johnson, D. A. Maltz and Y.C. Hu., “ The
Conference Proceeding, IEEE, Vol.1 21-24 Oct. 1996, dynamic source routing protocol for mobile ad hoc
pp. 225-229. networks (DSR),” IETF Mobile Ad hoc Networks
[11] P. Merlin and A. Segall, “A fail safe distributed routing Working Group, Internet Draft, 16 Apr. 2003.
protocol”, IEEE Transactions on Communications, [20] C. Bettstetter and C. Wagner. (March 2002), The
COM-27 (9), Sep. 1979, pp. 1280-1287. Spatial Node Distribution of the RandomWaypoint
[12] J.M. Jaffe and P.H. Moss, “A responsive distributed Mobility Model. In Proceedings of the German
routing algorithm for computer networks”, IEEE Workshop on Mobile Ad hoc Networks (WMAN),
Transactions on Communications, COM-30 (7): July pages 41–58, Ulm, Germany.
1982, pp. 1758-1762. [21]C. Bettstetter, (2001), “Mobility modeling in wireless
networks: Categorization, smooth movement, and

www.ijera.com 09


border effects,” ACM Mob. Comput. Commun. Rev., wireless networks,” in Proc. 4th ACM Int. Workshop
vol. 5, no. 3, pp. 55–66. on Modeling Analysis and Simulation of Wireless and
[22] C. Bettstetter, (Sept. 2001),“Smooth is better than Mobile Systems (MSWiM), Rome, pp. 19–27
sharp: a random mobility model for simulation of

www.ijera.com 010

011001010 a

More Related Content

What's hot (19)

Viewers also liked (20)

Similar to 011001010 a (20)

Recently uploaded (20)

011001010 a