International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
Ehsan Ahvar1, Alireza Pourmoslemi2, Mohammad Jalil Piran3
1
Department of Information Technology & Communication, Payame Noor University,
Iran
ehssana2000@yahoo.com
2
Department of Mathematics, Payame Noor University, Iran
pourmoslemy@yahoo.com
3
MIEEE, CSE Department, Jawaharlal Nehru Technological University, Hyderabad,
India
piran.mj@gmail.com , piran@ieee.org
ABSTRACT
Many energy-aware routing protocols have been proposed for wireless sensor networks. Most of them are
only energy savers and do not take care about energy balancing. The energy saver protocols try to
decrease the energy consumption of the network as a whole; however the energy manager protocols
balance the energy consumption in the network to avoid network partitioning. This means that energy saver
protocols are not necessarily energy balancing and vice versa. However, the lifetime of wireless sensor
network is strictly depending on energy consumption; therefore, energy management is an essential task to
be considered. This paper proposes an energy aware routing protocol, named FEAR, which considers
energy balancing and energy saving. It finds a fair trade-off between energy balancing and energy saving
by fuzzy set concept. FEAR routing protocol is simulated and evaluated by Glomosim simulator.
KEYWORDS
Sensor network; energy aware; routing protocol; Fuzzy Sets
1. INTRODUCTION
A new class of networks has appeared in the last few years: the so-called Wireless Sensor
Network (WSN). A WSN is a set of small autonomous systems, called sensor nodes (also known
as motes), which communicate wireless and cooperate to solve at least one common application.
These nodes have to collaborate to fulfill their tasks as usual a single node is incapable of doing
so. The sensor networks are included of many sensor nodes, which are deployed with high
density very close to the sensing environment or inside it, as the sensor environment is a field
where the sensors scattered to sense it and collect the required data. The tiny sensor nodes, which
consist of several major components as sensor, analog-to-digital convertor (ADC), processor,
storage, transmitter, power unit and some other components based on their task such as location
finding system, mobilizer and power generator, are able to interact with the phenomena and
collect data, and forward the received data as well as a router.
DOI: 10.5121/ijgca.2011.2203
32
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
Sensor nodes in the sensing field coordinate among themselves to produce high-quality
information about the physical environment. These sensors have the ability to communicate either
amongst each other or directly to an external base-station (BS). A base-station may be a fixed
node or a mobile node capable of connecting the sensor network to an existing communications
infrastructure or to the Internet where a user can access to the reported data. Many routing
algorithms have been developed for sensor and ad hoc networks [1-18, 28, 29]. All of these
routing protocols can be classified according to the network structure as flat, hierarchical, or
location-based. In flat networks, all nodes play the same role while hierarchical protocols aim at
clustering the nodes so that cluster heads can do some aggregation and reduction of data in order
to save energy. Location-based protocols utilize the position information to relay the data to the
desired regions rather than the whole network.
Energy consumption in sensor networks is a striking factor, it is because of that the batteries
carried by each mobile node have limited power supply, processing power is limited, which in
turn limits the quality and quantity of services and applications that can be supported by each
node. In sensor networks, the nodes play multiple roles of data providing, data processing and
data routing. The power required for data sensing varies to the kind of application. Energy
expenditure in data processing is much lower than data transmission, so power consumption is
the most considerable challenge in sensor networks. Routing in WSNs is very challenging due to
the inherent characteristics such as energy and bandwidth limitations that distinguish these
networks from other wireless networks like mobile ad hoc networks or cellular networks. This
paper addresses to design an energy-aware routing protocol for flat structure wireless sensor
networks. The proposed protocol, FEAR, tries to save and balance energy consumption in
network. It finds optimal route in energy level and hop-count both. Routing decisions in FEAR
are based on the distance to the Base Station as well as remaining energy of nodes on the path
towards the base station.
The rest of the paper is organized as follows, in section 2, we state some related works. Our
motivation is discussed, in section 3. In section 4, we introduce the fuzzy set. We present our
proposed methodology in section 5. Then we show the details and result of our simulation in
section 6 and 7, finally in section 7 we conclude our paper.
2. RELATED WORKS
In [23], a knowledge-based inference approach using fuzzy Petri nets is employed to select
cluster heads, and then the fuzzy reasoning mechanism is used to compute the degree of
reliability in the route sprouting tree from cluster heads to the base station. Finally, the most
reliable route among the cluster heads can be constructed. The algorithm not only balances the
energy load of each node but also provides global reliability for the whole network.
To overcome the limitations of LEACH, a fuzzy logic approach to cluster head election [24] is
proposed which uses three fuzzy variables (concentration, energy and centrality). However, this
algorithm is a centralized election mechanism, and the base station has to collect the energy and
distance information from all sensor nodes. In [25], cluster head election mechanism using fuzzy
logic (CHEF) is proposed, which is a localized cluster head election mechanism. CHEF uses
energy and local distance as fuzzy variables in the fuzzy if-then rules. Simulation results show
that the cluster heads in CHEF are more evenly distributed over the network than those in
33
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
LEACH, and then CHEF further prolongs the network lifetime. But CHEF does not construct
multi-hop routes in cluster heads.
A generalized fuzzy logic based energy-aware routing [26] is presented which is a soft, tuneable
parameter based algorithm. But this algorithm assumes that a cluster head is much powerful as
compared to the other sensor nodes and has no energy limitation. A fuzzy self-clustering
algorithm (FSCA) [27] considers the node residual energy and local density to improve the
lifetime of WSNs. To uniformly distribute clusters over the networks, FSCA employs migration
fuzzy module to re-cluster and merge existed clusters. However, re-clustering the whole network
adds more control overhead and needs more time.
3. MOTIVATION
Routing has a significant influence on the overall WSN lifetime, and providing an energy
efficient routing protocol remains an open research issue [9]. Most of energy aware routing
protocols are designed to save total energy consumption. They usually find the shortest path
between Source and Sink to reduce energy consumption. In our opinion, an energy saver protocol
that balances energy consumption is better than a poor energy server protocol. To find the best
route with respect to the "shortest path" may lead to network partitioning. On the other hand, if
the attention is only paid to the energy balancing, may cause the found route is a long path and
the network lifetime decreased as well. On the other hand, finding best route only based on
energy balancing consideration may lead to long path with high delay and decreases network
lifetime. [22]
The SEER is a simple energy efficient routing protocol [11]. It tries to reduce number of
transmissions. But it has poor idea about energy management and energy balancing. On the other
hand, the LABER routing protocol [1] tries to balance energy consumption. But it has some
impressive problems such as low accuracy in updating of energy and high control overhead. In
the LABER, the Acknowledgement packet only is forwarded to previous sender, but the other
neighbors cannot update energy level of the sender of Acknowledgement; low accuracy.
Moreover, the Acknowledgment packet is an extra control overhead. This paper designs an
energy-aware routing protocol for Wireless Sensor Networks to balance and save energy by fuzzy
set technique. It also updates energy without Acknowledgment packets to increase accuracy and
decrease control overhead.
4. INTRODUCTION TO FUZZY SET
The theory of fuzzy sets was introduced by Prof. L. Zadeh in 1965 [19]. After the pioneering
work of Prof. Zadeh, there has been a great effort to obtain fuzzy analogues of classical theories.
Fuzzy set theory is a powerful tool for modeling uncertainty and for processing vague or
subjective information in mathematical models. Their main directions of development have been
diverse and its applications to the very varied real problems. The notion central to fuzzy systems
is that truth values (in fuzzy logic) or membership values (in fuzzy sets) are indicated by a value
on the range [0, 1], with "0" and "1" representing absolute Falseness and absolute Truth
respectively. Some concepts of fuzzy set theory are listed as follows [21]:
34
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
Definition 3-1: Let X be some set of objects, with elements noted as x.
Definition 3-2: A fuzzy set A in X is characterized by a membership function mA(x) which maps
each point in X onto the real interval [0,1]. As mA(x) approaches 1, the "grade of membership"
of x in A increases.
Definition 3-3: A is EMPTY iff for all x, mA(x) =0
Definition 3-4: A = B iff for all x: mA(x) = mB(x)
Definition 3-5: mA' = 1 - mA.
Definition 3-6: A is CONTAINED in B iff mA
mB.
Definition 3-7: C = A UNION B, where: mC(x) = MAX(mA(x), mB(x)).
Definition 3-8: C = A INTERSECTION B where: mC(x) = MIN(mA(x), mB(x)).
Definition 3-9: Given a fuzzy set A , the alpha-cut (or lambda cut) set of A is defined by
~
~
Aα = {x mA(x) ≥ α }
5. FEAR ROUTING PROTOCOL
FEAR protocol is a reactive protocol which employs a lazy approach whereby nodes only
discover routes to destination only when needed (on-demand). FEAR consumes much less
bandwidth than proactive protocols such as Destination-Sequenced Distance-Vector (DSDV)
protocol, but the delay in determining a route can be substantially large. FEAR protocol is an
energy aware routing protocol, which consists of three major steps:
•
Neighbor discovery
•
Forwarding data
•
Energy Update
Below we explain the steps in details;
5.1 Neighbor discovery
The Sink or the Base Station (BS) initializes the network by flooding the network with a
broadcast message. Each node that receives the initiate packet, adds an entry to its “Neighbor
Table” including “Neighbor ID”, “Energy Level” and “Hop Count” to its “Neighbor Table”.
Then, the node make some changes on broadcast message such as (1) it increments the “Hop
Count” field, (2) it changes the “Source Address” field to its address and (3) it changes the
“Energy Level” field to its energy level and then retransmits it. Every node in the network
retransmits the broadcast message only once, to all of its neighbors. When the initial broadcast
message has been flooded through the network, each node knows hop count and energy level of
its neighbors.
The Sink node periodically sends a broadcast message through the network; so that nodes add
new neighbors joined the network to the “Neighbor Table” and remove neighbors that have failed
to be an active member of the network.
35
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
5.2 Forwarding data
When a node observes an event, it initiates a routing process. One of the most challenges in the
reactive protocols is how to select the next hop. However, in this paper we propose a new
scenario to solve it. The proposed protocol, FEAR, is based on fuzzy set technique. It consists of
two fuzzy sets; A and B.
A is the fuzzy set of all neighbors’ energy levels:
A = { e1 ,e2 ,...,en }
A has a membership function, mA( ei ) which can be defined as below:
mA(ei ) = λei , 1 ≤ i ≤ n .
(1)
where is a control parameter to limit energy factor in [0,1] interval and
ei
is energy level of (i)th
neighbor.
Let
be obtained from the following formula:
n
mA( ei )
α=
where
i =1
Then
n
Aα = {ei mA( ei ) ≥ α }
(2)
α is energy threshold and Aα ( α -cut) is used to remove the neighbors with unacceptable energy
level.
and B is the fuzzy set of all neighbors’ hop counts with membership function mB( hi ).
B = { h1 , h2 ,...,hn }
mB( hi ) = 1 −
hi
MaxHop
Where n is the number of neighbors, and
hi
,
1≤i ≤ n
(3)
is hop count of (i)th neighbor.
Now, we define following decision maker equation:
C( i ) =
mAα ( ei ) × mB( hi )
0
λei > α
, Where 1 ≤ i ≤ n .
λei ≤ α
(4)
However, the neighbor with maximum amount of C is selected as the next hop.
As used in equation (3), “MaxHop” is the estimation of the longest possible route in the network
which plays an essential role on the FEAR's decisions. This paper tries to select optimal
"MaxHop". Very large amount of MaxHop increases effect of "hopcount" factor compared with
energy on equation (1) and vice versa.
36
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
To compute MaxHop some methods are proposed and discussed in this paper. These methods are
classified into two different categories: Dynamic and static methods. For all proposed methods
the sensor network with size of X*Y and nodes with Radio range of R is supposed.
First method computes maximum hop count based on the longest available distance and radio
range of R, Fig.1 (a). It finds maximum possible distance between two nodes and then computes
the “MaxHop” based on the following equation:
MaxHop = ( X 2 ) + (Y 2 ) / R
(5)
However, the previous method considers the longest available path between Source and
Destination and computes maximum hop count based on longest available distance. But the
distance between two nodes or range of one hop was assumed with fix rang of R. This
assumption may not be feasible. There is no accrued knowledge about the distance between two
nodes.
To obtain an accurate “MaxHop”, R is changed to R/2, Fig.1 (d), therefore:
MaxHop = ( X 2 ) + (Y 2 ) /( R / 2)
(6)
Also consider a network consisting of N nodes. It is possible to have a path between source and
destination with N-2 hops Fig.1 (b). In this condition the “MaxHop = N-2”.
Figure 1 : Different shapes of finding MaxHop
Generally, static “MaxHop” method is not too accurate. “MaxHop” is difficult to find and also
constant amount of “MaxHop” leads to increase effect of “hopcount” factor for the nodes close to
the sink and decrease effect of this factor for the further nodes.
37
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
Assume all nodes have equal energy levels.
Nodes which are close to the Sink,
HopCount<<MaxHop and Hop factor is almost one. But for the nodes located in a long distance
to the sink, HopCount~MaxHop and Hop Factor will be about zero. Therefore, at different
locations of network the effect of Hop factor is variable and constant “MaxHop” can not support
a fair tradeoff between Hop and Energy Factors, Fig.2.
Figure 2 : Static method scheme
Because of weaknesses of static methods, the Dynamic Method is proposed Fig.1 (c). In this
method “MaxHop” is computed prior to each forwarding step. In this scenario, each node looks
at its ‘Neighbor Table” and finds out the biggest hop count. Then; MaxHop = Biggest hop count
+ 1.
5.3 Energy Update
All the neighbors of the sender node receive the forwarded data packet via the overhearing
technique. Then, they update the energy level of sender node in their "Neighbor Table" by the
piggybacking technique. Nodes might be used by more than one neighbor for routing; in this case
the energy value stored in the "Neighbor Tables" of both the node's neighbors will be completely
accurate by the overhearing technique.
The operation of FEAR protocol can be summarized as follows:
[0] The sink initializes the network by flooding the network with a broadcast message.
[1] Nodes add all their neighbors’ information to their neighbor tables.
[2] The node with highest C(i) is selected and data packet is forwarded to it.
[3] The energy level of data sender node is updated by its neighbors.
38
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
FEAR routing protocol does not need any energy message because the energy level of each
sender is updated into its "Neighbor Table" by overhearing and piggybacking techniques
automatically. Also, the sink node sends a broadcast message through the network, so that nodes
can add new neighbors that joined the network to the "Neighbor Table" and remove neighbors
that have failed. However, the sending rate of broadcast message through the network is related to
the nodes’ mobility. In contrary to the networks with high mobility nodes, the networks with non
mobility nodes do not need to send broadcast message.
6. SIMULATION METHODOLOGY
For simulation work two features of the proposed protocol should be considered: (1) FEAR with
dynamic “MaxHop”, known as Dynamic FEAR (D-FEAR) and (2) FEAR with Constant
“MaxHop”, known as Static FEAR (S-FEAR). The constant “MaxHop” is obtained based on radio
range of R/2, and if maximum hop count is bigger than “MaxHop”, then maximum hop count will
be assumed “MaxHop”. To compare the routing protocols, a parallel discrete event-driven
simulator, GloMoSim, is used. GloMoSim is a simulation tool for large wireless and wired
networks [20].
Table 1 describes the detailed setup for our simulator.
Table 1.
Simulation Setting
SIMULATION-TIME
1200 SECOND
TERRAIN-DIMENSIONS
1000m*1000m
NUMBER-OF-NODES
200, 300, 500, 1000, 2000
NODE-PLACEMENT
Uniform/Random
MOBILITY
NUMBER OF EVENTS (Sources)
TEMPARATURE
RADIO-BANDWIDTH
RADIO-TX-POWER
ENERGY- TRANSMITLEVEL
MAC-PROTOCOL
NETWORK-PROTOCOL
PROPAGATION-PATHLOSS
RADIO-TYPE
NONE
100
290.0 (in K)
2000000(in bps)
5.0 (in dBm)
0.0002 (in mW)
802.11
IP
FREE-SPACE
RADIO-ACCNOISE
39
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
7.
SIMULATION RESULT
In this section we evaluate and compare the various routing schemes. The interested performance
measures in this study are: (a) Average energy consumption of transmission (in mW); (b) Energy
balancing. The variables are: number of nodes and node placement.
The simulation of the protocol started with a broadcast message. We select 100 Sources to send
data packets to the Base Station during the simulation. The Sources are selected randomly in
different times. Each Source generates a 512-bit data packet and forwards it through the network.
Simulations are performed to evaluate the network lifetime achieved by each protocol. At the
beginning of simulation, the transmission energy level of each node was 0.0002 mW. We design
four different tests to evaluate protocols as follows:
Test 1: The time until the first neighbor of station fails. The first failure of station neighbor is
related to energy management. Energy balancer protocols should have better failure time than the
other protocols. Therefore, this test evaluates energy management of each protocol.
Test 2: The Number of fails. This test computes number of fails because of their depleting their
energy sources. The protocol with the lowest number of fails is the best in energy factor.
Test 3: Percentage of active neighbors of the station at the end of simulation. This test shows
protocols’ ability to keep the station connected.
Test 4: The average remaining energy of all the nodes in the network, at transmission mode. This
test is designed to find which protocol is more energy saver than the other.
After the simulation the following results are achieved:
After the simulation the following results are achieved:
SEER
D-FEAR
S-FEAR
1400
1200
Time[s]
1000
800
600
400
200
0
200
300
500
Nodes
1000
2000
Figure 3: Time at which the first neighbor of Station fails due to
depleting its energy source. The nodes are randomly distributed over
the sensor area.
40
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
SEER
D-FEAR
S-FEAR
1400
1200
Time[s]
1000
800
600
400
200
0
200
300
500
Nodes
1000
2000
Figure 4 : Time at which the first neighbor of Station fails due to
depleting its energy source. The nodes are uniformly distributed over
the sensor area.
SEER
160
140
D-FEAR
120
S-FEAR
Fails
100
80
60
40
20
0
200
300
500
Nodes
1000
2000
Fails
Figure 5 : Number of fails at the end of simulation. The nodes are
randomly distributed over the sensor area.
300
SEER
250
D-FEAR
200
S-FEAR
150
100
50
0
200
300
500
1000
2000
Nodes
Figure 6 : Number of fails at the end of simulation. The nodes are
uniformly distributed over the sensor area.
41
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
SEER
D-FEAR
S-FEAR
120
Actives(%)
100
80
60
40
20
0
200
300
500
Nodes
.
1000
2000
Figure 7 : Percentage of active neighbors of Station at the end of
simulation. The nodes are randomly distributed over the sensor area.
SEER
D-FEAR
S-FEAR
120
Actives(%)
100
80
60
40
20
0
200
300
500
Nodes
1000
2000
Avg En Consumption(mW)
Figure 8 : Percentage of active neighbors of Station at the end of
simulation. The nodes are uniformly distributed over the sensor area.
SEER
0.00025
D-FEAR
0.0002
S-FEAR
0.00015
0.0001
0.00005
0
200
300
500
Nodes
1000
2000
Figure 9 : Average energy consumption (mW) in transmission mode.
The nodes are randomly distributed over the sensor area.
42
Avg En Consumption(mW)
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
0.00025
SEER
D-FEAR
0.0002
S-FEAR
0.00015
0.0001
0.00005
0
200
300
500
Nodes
1000
2000
Figure 10 : Average energy consumption (mW) in transmission mode.
The nodes are uniformly distributed over the sensor area.
Results of Test 1 (figure3 and figure 4) show that there is a considerable discrepancies between
the SEER and FEAR protocols in the time at which the first neighbor of the Base Station fails due
to its depleting energy source. D-FEAR and S-FEAR are more optimal than SEER protocol in
energy management. This is due to the fact that FEAR sends data packet along an optimum
energy-balanced path. But there is no considerable difference between D-FEAR and S-FEAR
Results of Test 2 (figures 5 & 6) show that at the end of simulation, FEAR has very low fails
compared with the SEER. Consequently, our proposed protocol has an acceptable performance in
high-density networks.
Test 3 (figures 7 & 8) shows the percentage of active neighbors of the Base Station at the end of
simulation. One can see that D-FEAR and S-FEAR have better performance than SEER
especially in high density networks.
Test 4 (figures 9 & 10) shows that there is no visible difference in energy consumption between
SEER, D-FEAR and S-FEAR.
As yielded desired and acceptable results by the abovementioned four tests, it is salient that
FEAR has fair performance in “Energy Balance” aspect. So, it’s more suitable for wireless sensor
networks. According to the test results it is obvious that the proposed FEAR routing protocol
increase the network lifetime in compare to the SEER. Also, the performance of the S-FEAR has
been improved by presenting D-FEAR.
8.
CONCLUSION
In this paper a protocol for routing with respect to energy saving and balancing was proposed for
wireless sensor networks. In our proposed protocol, FEAR, we utilized fuzzy techniques in order
to achieve energy balancing. We have simulated our job with Golomosim Simulator. The FEAR
was compared to traditional FEAR. As observed FEAR in energy balancing has the better
performance than SEER especially in high density networks i.e. wireless sensor networks.
However, we have seen that FEAR showed low failures too. Finally, we conclude that FEAR is a
43
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
good proposal for wireless sensor networks in subject of routing with respect to its proved
performances in energy consumption as well as energy balancing.
REFERENCES
[1]
S. M. Abolhasani, M. R, Meybod, M. Esnaashari, "LABER: A Learning Automata Based Energyaware Routing Protocol for Sensor Networks", Proceedings of the Third Information and
Knowledge Technology conference, Ferdowsi University, Iran, Nov. 27-29, 2007.
[2]
E.Ahvar, M.Fathy, “BEAR: A Balanced Energy-Aware Routing Protocol for Wireless Sensor
Networks” Wireless Sensor Network Journal, vol. 2, pp. 768–776. October 2010.
[3]
T.Banka, G.Tandon, A.P. Jayasumana "Zonal Rumor Routing for Wireless Sensor Networks"
Proceedings of IEEE Conference on Information Technology: Coding and Computing, ITCC
2005, pp. 562-567, Las Vegas, NV, April 2005.
[4]
F. Block, C. Baum. “An energy-efficient routing protocol for wireless sensor networks with
battery level uncertainty”. In Proceedings IEEE Military Communications Conference, vol. 1, pp.
489–494. 7–10 October 2002.
[5]
D. Braginsky, D.Estrin, “Rumor Routing Algorithm for Sensor Networks" In Proc. First ACM
Workshop on Sensor Networks and Applications, Atlanta, GA, USA, pp. 22–31R. October 2002,
[6]
C. Chen. “Energy efficient routing for clustered wireless sensors network”. In Proceedings 29th
Annual Conference of the IEEE Industrial Electronics Society, vol. 2, pp. 1437– 1440. 2–6
November 2003.
[7]
J. Chen, Y. Hsu and I. Chang, “Adaptive Routing Protocol for Reliable Sensor Network
Applications”, International Journal on Smart Sensing and Intelligent Systems, Vol. 2, No. 4,
December 2009.
[8]
A. Datta. “Fault-tolerant and energy-efficient permutation routing protocol for wireless
networks”. In Proceedings International Parallel and Distributed Processing Symposium. 22–26
April 2003.
[9]
S.Fedor, “Cross-Layer Energy Optimization of Routing Protocols in Wireless Sensor Networks”
PhD thesis, Dublin City University, April 2008.
[10]
M.Gholipour, M.R.Meybodi, "LA-Mobicast: A learning automata based distributed protocol for
mobicast in sensor networks" Proc of CSICC2007, PP.1154-1161, 2007.
[11]
G.P. Hancke, C.J. Leuschner, “SEER: A Simple Energy Efficient Protocol for Wireless Sensor
Networks” South African Computer Journal (SACJ), No.39, 2007.
[12]
T. Hou, Y. Shi and J. Pan. “A lifetime-aware single-session flow routing algorithm for energyconstrained wireless sensor networks”. In Proceedings IEEE Military Communications
Conference, vol. 1, pp. 603–608. 13–16 October 2003.
[13]
C. Intanagonwiwat, R. Govindan and D. Estrin. “Directed diffusion: A scalable and robust
communication paradigm for sensor networks”. pp. 56–67. August 2000.
[14]
Y.M.Luz, V. W. S. Wong, “An Energy-Efficient Multi Path Routing Protocol for Wireless Sensor
Networks”, International Journal of Communication System, pp 747-766, 2007.
[15]
A.Manjeshwar and D. P. Agrawal. “APTEEN: a hybrid protocol for efficient routing and
comprehensive information retrieval in wireless sensor networks”. In IPDPS ’02: Proceedings of
the 16th International Parallel and Distributed Processing Symposium, p. 48. 15–19 April 2002.
44
International Journal of Grid Computing & Applications (IJGCA) Vol.2, No.2, June 2011
[16]
C. Schurgers and M. Srivastava. “Energy efficient routing in wireless sensor networks”. In
Proceedings IEEE Military Communications Conference, vol. 1, pp. 357–361. 28–31 October
2001.
[17]
M.Singh, M.Sethi, N.Lal and S.Poonia “A Tree Based Routing Protocol for Mobile Sensor
Networks” International Journal on Computer Science and Engineering, Vol.2, No.1s, pp 55-60,
2010.
[18]
D. Tian and N. Georganas. “Energy efficient routing with guaranteed delivery in wireless sensor
networks”. In Proceedings IEEE Wireless Communications and Networking Conference, vol. 3,
pp. 1923–1929. 16–20 March 2003.
[19]
L.A.Zadeh, "Fuzzy sets". Inform. Control 8, 338–353, 1965
[20]
X.Zeng, R.Bagrodia, and Geria M. "Glomosim: A Library for Parallel Simulation of Large Scale
Wireless Networks" in Proceedings of the 12th Workshop on Parallel and Distributed Simulations,
154-161, 1998.
[21]
H.J. Zimmermann, "Fuzzy Set Theory and its Applications", third ed. Kluwer Academic
Publishers, Boston, MA. 1996.
[22]
Ehsan Ahvar, Mohammad Jalil Piran, “SELECTOR: An Intelligent Evaluation System for
Routing Protocols in Wireless Ad Hoc and Sensor Networks”, Proceeding of 3rd International
Conference on Electronics Computer Technology, ICECT 2011, IEEE, Kanyakumari, India, p.p.
300~305, April 8~10, 2011.
[23]
1. Zhenhua Yu, Xiao Fu , Yuanli Cai and Mehmet C. Vuran "A Reliable Energy-Efficient MultiLevel Routing Algorithm for Wireless Sensor Networks Using Fuzzy Petri Nets"Sensors 2011
[24]
Gupta, I.; Riordan, D.; Sampalli, S. Cluster-head election using fuzzy logic for wireless sensor
networks. In Proceedings of the 3rd Annual Communication Networks and Services Research
Conference, Nova Scotia, Canada, 16–18 May 2005; pp. 255-260.
[25]
Kim, J.M.; Park, S.H.; Han, Y.J.; Chung, T.M.; CHEF: Cluster head election mechanism using
fuzzy logic in wireless sensor networks. In Proceedings of International Conference on Advanced
Communication Technology, Dublin, Ireland, 17–20 February 2008; pp. 654-659.
[26]
Haider, T.; Yusuf, M. A fuzzy approach to energy optimized routing for wireless sensor networks.
IAJIT 2009, 6, 179-188.
[27]
Tashtoush, M.; Okour, M.A. Fuzzy self-clustering for wireless sensor networks. In Proceedings of
IEEE/IFIP International Conference on Embedded and Ubiquitous Computing, Shanghai, China,
17–20 December 2008; pp. 223-229.
[28]
Mohammd Jalil Piran, G. Ramma Murthy, "A Novel Routing Algorithm for Vehicular Sensor
Networks", International Journal of Wireless Sensor Networks, USA, pp. 919-923, December
2010.
[29]
Mohammd Jalil Piran, G. Ramma Murthy, G. Praveen Babu "A New Approach to Vehicular
Sensor Networks", International Journal of Wireless Networks and Communications, p.p. 1~12,
June 2011.
45