Enhanced Passive Clustering Based on Distance and Residual Energy for Wireless Sensor Network

Abderrahim MaizateSTIC Laboratory

Chouaib Doukkali University, B.P: 20, El Jadida, Morocco

abderrahim.maizate@gmail.com

Najib El kamounSTIC Laboratory

Chouaib Doukkali University, B.P: 20, El Jadida, Morocco

elkamoun@gmail.com

Abstract— Wireless sensor network consists of a set of tiny sensor nodes. The nodes are continuously sense and transmit the data to the base station. Therefore, energy consumption and network coverage are important issues to improve the life time of the network. Clustering architecture is an effective architecture toreduce the energy consumption. The main aim is to increase network lifetime as well as increase the presence of live nodes so that more nodes will remain exist. In this paper we propose a new approach called EPCDRE (Enhanced passive clustering algorithm based on distance and residual energy), which evenly distributes the energy dissipation among the sensor nodes tomaximize the network lifetime. This is achieved by using residualenergy and distance between nodes in the selection of nodesclusterheads and election of clusterhead backup. Comparisonwith the existing schemes such as Passive Clustering, PCEEC and GRIDS (Geographically Repulsive Insomnious DistributedSensors) reveals that the proposed algorithm approachsignificantly improves the network lifetime and it can further efficiently relay the cluster data.

Keywords-wireless sensor networks, self-organization,Clustering passive, clustering, network lifetime, energy efficiency, fault tolerance, Residual Energy.

I. INTRODUCTION

Wireless Sensor Networks (WSN) is emerging as a rich domain of active research involving hardware, networking, distributed algorithms and other disciplines. The sensor nodeshave the limited energy, are mostly set in the area where is dangerous or not easily accessible [1]. Accordingly, increasing the lifetime of the network and preserve the energy of the different nodes is a challenge for these networks. Manydifferent self organization protocols have been designed for wireless sensor networks (WSNs), dependent on both thearchitecture of wireless sensor network (WSN) and theapplications that WSN is intended to support. These protocols support requirements and constraints of WSNs and efficiently make them an integral part of recent technologies. Figure 1[19]shows the structure of a wireless sensor network.

Consequently; several academic and industrial applications have been developed based on wireless sensor networks. Theseapplications cover many aspects ranging from military tomedical applications. The idea beyond these applications is that; densely deploying sensor nodes with capabilities of

Figure 1. General Sensor Network Architecture

sensing, wireless communications, and computation in an unattended environment and will assist in measuring ambient conditions in that specific environment, and obtaining thecharacteristics about the physical phenomenon surrounding these sensors; by transmitting these sensed and gathered data to the base station. The gathering of data is required in order to decrease the waste of energy caused by the exchange of databetween the adjacent sensors nodes, clustering techniques have been proposed among the best solutions. These techniques organize the nodes into clusters where some nodes work as clusterheads and collect the data from other nodes in the clusters. Then, the clusterheads can consolidate the data and send it to the base station as a single packet, thus reducing thenumber of packets exchanged in the network.

The WSNs protocols are different from conventional ones; in essence they need to support various unique requirements and constraints to make wireless sensor networks practically useful and operating, these requirements and constraints areintroduced by factors such as: self-organizing, small-size, low-power consumption, fault-tolerance, low-latency, scalability,adaptivity, and robustness. In this paper, we present a new algorithm for self organization based clustering and a review of some ongoing work on designing and developing selforganization protocols for wireless sensor networks. Inaddition, we describe their advantages and deficiencies.

978-1-4799-0792-2/13/$31.00 ©2013 IEEE

In a clustered network, the nodes are grouped into clustersaround a clusterhead. The energy efficiency of a clustered sensor network depends on the selection of the Clusterheads.Heinzelman et al. [2] propose a low-energy adaptive clustering hierarchy (LEACH), which generates clusters based on the size of the sensor network. However, this approach needs a priori knowledge of the network topology. Younis and Fahmy [3]propose a Hybrid Energy-Efficient Distributed clustering(HEED), which creates distributed clusters without the size and density of the sensor network being known. However, the cluster topology fails to achieve minimum energy consumption in intra-cluster communication. In [4], they propose an energy-aware cluster formation protocol (GRIDS) which increase the lifespan of a sensor network by using an efficient selection mechanism of critical (or not) nodes. It promotes energyefficiency by reducing communication and effective sleepmode operation. By examining remaining power level, low powered sensor nodes can go to sleep for the next round of operation period. GRIDS inherits many advantages fromPassive Clustering. In [5], we define a protocol for clusterformation and election of clusterheads (PCEEC) whichprovides several advantages. It uses balanced energyconsumption among network nodes, minimizes the number ofclusters (Clusterhead) and provides effective coverage of thenetwork, thus it keeps longer the structure of clusters andminimize the consumed energy. As a result, the networkstability is preserved and the lifetime of the network issignificantly increased.

The rest of the paper is organized as follows. In Section 2, we present related works. Section 3 describes both the energy model and network model. Section 4 describes the proposed algorithm. Experimental results are reported in Section 5 andconclusions are drawn in Section 6.

II. RELATED WORK

Depending on the structure of the network, there are several categories of self organization protocols based on clustering:grid schemes, heuristic schemes, weighted schemes andhierarchical schemes. Clustering is an important research topic in the areas of wireless sensor network (WSN) becauseclustering improves the performance of many systems. In WSN, clustering can be used to improve the networkperformance through quality of service metrics such asthroughput and delay, in the presence of both mobility and a large number of mobile nodes with minimal resources. Thissection provides the brief description about the clusteringtechnique, its objectives and role.

Passive clustering [6] can be described as on demandcluster formation protocol that does not use dedicated protocol-specific control packets or signals. The formation of cluster is dynamic and initiated by the first data message to be flooded. Which in turn reduces the duration of the initial set-up period,and the benefits of the reduction of the forwarding set can be felt by calculating the total energy consumed because the main function of the clusters is to optimize the exchange of flooded messages. Each node collects neighbor information when there are on-going data packets and can construct clusters even without collecting the complete neighbor list (figure 2). This is an innovative approach to clustering which virtually eliminates

major cluster overheads, the time latency for initial clustering construction as well as the communication overhead forneighbor information exchanges. Instead of using protocolspecific signals or packets, cluster status information (2 bits for four states: Initial, Clusterhead, Gateway, and Ordinary-nodestates) of a sender is stamped in a reserved field in the packet header. Sender ID (another key piece of information forclustering) is carried by all the existing MAC protocols and can be retrieved from the MAC header. Since in flooding the MAC packets are transmitted in broadcast (instead of unicast) mo de,every node receives and reads the packets (in a promiscuous way), and thus participates in passive clustering.

Figure 2. Cluster Structure with Passive Clustering (PC)

In passive clustering each node operates the MAC sender address carried by the received packets to collect neighborinformation, and can construct clusters even without collecting the complete neighbor list. Instead of using protocol specific signals or packets, passive clustering reserves two bits for the following four states of a mobile node: ( 1) Initial, (2) Cluster head, (3) Gateway and (4) Ordinary. At the beginning, every sensor node is in the INITIAL state until it receives a packet. If the sender’s packet is not CLUSTERHEAD, this sensor node switches to CLUSTERHEAD-READY. This node will becomea CLUSTERHEAD if it successfully transmits a packet before receiving any packets from others. If the sensor node receives a packet from a CLUSTERHEAD it changes state toORDINADRY. Any sensor node that hears more then oneCLUSTERHEAD becomes GATEWAY.

Passive clustering has several mechanisms for the cluster formation such as: Gateway Selection Heuristic and First Declaration Wins rule. The Gateway Selection Heuristicprovides a procedure to elect the minimal number of gateways required to maintain the connectivity between clusterheads.With the First Declaration Wins rule, a node that first transmits a data message will be a clusterhead of the rest of nodes in its clustered area.

Passive clustering maintains clusters using implicit timeout. A node assumes that some nodes are out of clustered area if they have not sent any data longer than timeout duration. With reasonable offered load, a node can catch dynamic topology changes. The advantages of this approach can be summarized as:

• Clustering can be achieved without using protocol specific, explicit control packets or signals.

• Passive clustering does not require the initial clustering phase to precede the data andcommunication phase.

• Passive clustering does not require re-clustering to satisfy clustering regulation, when theconnectivity changes, because of mobility.

• Clustering can be done without collectingcomplete neighborhood information.

GRIDS [4] is an energy-aware cluster formation protocol which increase the lifespan of a sensor network by using anefficient selection mechanism of critical (or not) nodes. This mechanism allows balanced energy consumption among the sensor nodes without requiring additional overheads including additional signaling, time synchronization and globalinformation. GRIDS is based on an energy model which delivers node’s residual/remaining energy level in real time. This information is piggybacked in the nodes packet header. Each sensor determines being insomnious or not based on its residual energy and the number of neighbouring insomnious nodes and their energy level. An efficient flooding during each wake up period determines insomnious nodes in the network. GRIDS selects insomnious nodes well distributed in the sensor deployed area. GRIDS inherits PC for constructing andmaintaining clusters. The main differentiator is that a set of nodes in a cluster with higher energy levels have higherprobability to become critical nodes, i.e., CH or GW. In PC, CHs keep their cluster status until there is a CH collision, i.e. the hop distance between two CHs becomes 1, and one of them resigns from CH. In GRIDS, an energy abundant node can challenge CH and usurps the role. Even if there is a CH declaration, nodes can challenge when their energy levels are higher than the one of CH. These nodes keep their cluster status even if they receive packets from the current CH.

In GRIDS, when a source node sends the first data packet, all of the nodes in its radio range become Clusterehead _ready. The first node that succeeds forwarding the packet becomes a Clusterehead . All nodes in the cluster will add 1 to the number of clusterhead only if the residual energy level of theClusterehead is higher than its own energy level. Comparing the number of clusterhead with higher energy level and the number of gatway, non-Clusterehead nodes determine theclustering status. When a gateway with a higher energy level sends a data packet, each recipient adds 1 to the number of gateway and some nodes with lower energy level becomeordinary nodes. When there is another data sink and if the initialization flooding is initiated at the same time from a different region, the insomnious node selection can beexpedited and the resulting insomnious node selection might be more efficient, i.e., smaller number of insomnious sensors. When a node has lower energy level compared with ones of neighbouring nodes, it is most likely that its cluster status becomes ordinary nodes and the node will sleep for this duty cycle. When a node reaches its lower energy level, GRIDS allows it to finish its transaction and declares it dead. By this way, GRIDS balances energy consumption by network nodes.

PCEEC (Passive Clustering for Efficient EnergyConservation in Wireless Sensor Network) [5] defines aprotocol for cluster formation and election of alternates of the

clusterheads. The clusterhead election is based on the energy and distance between nodes. PCEEC allows the election of an alternate for each cluster head and a dynamic balancing of the role of clusterhead to the alternate when leaving or failure. Thus, it provides several advantages network reliability,stability of clusters and reduces energy consumption among the sensor nodes. PCEEC uses the same principles as passiveclustering for the construction and maintenance of clusters inwireless sensor networks. It also inherits the characteristics of the algorithm GRIDS [4] by giving nodes with the highest level of energy to become a critical node, i.e., ClusterHead,Alternate or GateWay.

LEACH (low energy adaptive clustering) [2] is the most well known energy-efficient clustering protocol for WSNs that uses coordination in the clustering process. In LEACH the nodes organize themselves into local clusters, with one node acting as a cluster-head and exploiting data aggregation in the routing protocol to reduce the amount of data packet that must be transmitted to the base station. The aggregation is assumed to be perfect. The cluster head is assumed to aggregate the packets received so that one packet is sent, irrespective of the number of packets received. The idea behind LEACH is to form clusters of the sensor nodes depending on the received signal strength and use local cluster heads as routers to route data to the base station. The key features of LEACH are:

• Localized coordination and control for cluster set-up and operation.

• Randomized rotation of the cluster”base stations” or”cluster-heads” and the corresponding clusters.

• Local compression to reduce globalcommunication.

The operation of LEACH is divided into rounds. Each round begins with a set-up phase when the clusters areorganized, followed by a steady-state phase when data aretransferred from the nodes to the base station. During the set-upphase, a sensor node chooses a random number between 0 and 1. If this random number is less than the threshold T (n), the sensor node become a cluster-head for the current round. In steady state phase data transfer is done in a single hop manner to the base station. In each cluster one node is selected as a cluster head and the rest of node’s clusters are called cluster member. Data collected from member nodes are processed in cluster head before sending to a base station, and then will be sent to the base station after aggregation of data in the form of a package. Figure 3 showed the basic topology of LEACH.

Figure 3. The Basic Topology of LEACH.

Although, LEACH has shown good features to sensornetworks, such as clustering architecture, localizedcoordination and control, randomized rotation of cluster head, and local compression to reduce global communications(energy consumption minimization). However, there exist afew disadvantages in LEACH as follows:

• It performs the single-hop inter-cluster, directly from CHs to the base station which is notapplicable to large-region networks.

• Despite the fact that CHs rotation is performed at each round to achieve load balancing, LEACH cannot ensure real load balancing in the case of sensor nodes with different amounts of initial energy, because CHs are elected in terms of probabilities without energy considerations.

• The CHs are not uniformly distributed throughout the network.

• The idea of dynamic clustering brings extraoverhead.

LEACH is a completely distributed approach and requires no global information of network. In the literature, various improvements have been made to the LEACH protocol, which form LEACH family, such as TL-LEACH [11], E-LEACH[12], M-LEACH [13], LEACH-C [14], V-LEACH [15],LEACH-FL [16], W-LEACH [17], T-LEACH [18], etc.

In HEED protocol, residual energy of each sensor node is the primary parameter for probabilistic election of cluster-heads [3]. There are four primary goals of HEED. These are listed below:

• Prolonging the lifetime of the wireless sensornetwork by evenly distributing energyconsumption method that benefits from the use of the two important parameters for CH election;

• Selecting cluster-heads in a constant number of iterations and provides uniform CH distribution across the network and load balancing;

• Minimization of control overhead

• Formation of well-distributed cluster-heads and compact clusters

• Communications in a multi-hop fashion between CHs and the BS promote more energyconservation and scalability

In HEED, CHs are periodically elected based on twoimportant parameters: residual energy and intra-clustercommunication cost of the candidate nodes. Initially, in HEED, a percentage of CHs among all nodes, Cprob, is set to assume that an optimal percentage cannot be computed a priori. The probability that a node becomes a CH is:

(1)

where EResidual is the estimated rema ining energy of the node, and Emax is a reference maximum energy, which is typically identical for all nodes in the network. In HEED, elected CHs have relatively high average residual energy compared toordinary nodes . Additionally; one of the main goals of HEED is to get even distributed CHs throughout the networks.Moreover, despite the phenomena that two nodes, within each other’s communication range, become CHs together, but the probability of this phenomena is very small in HEED.

However, there are some limitations with HEED asfollows:

• Similar to LEACH, the performing of clustering in each round imposes significant overhead in the network. This overhead causes noticeable energy dissipation which results in decreasing thenetwork lifetime;

• HEED suffers from a consequent overhead since it needs several iterations to form clusters. At eachiteration, a lot of packets are broadcast.

• Some CHs, especially near the sink, may dieearlier because these CHs have more work load, and the hot spot will come into being in the network

Compared with self organization protocols in wirelesssensor networks, clustering protocols have a variety ofadvantages, such as more scalability, less load, less energy consumption and more robustness.

III. ENERGY AND NETWORK MODEL

In this section, we present the model of energy that will be used in the performance evaluations section, assumptions made herewill be used in the performance evaluation section.

A. Energy modelThe energy model used is same with that in Ref. [7], shown

in Figure 4. Equation (1) represents the amount of energy consumed for transmitting l bits of data to d distance. Equation (2) represents the amount of energy consumed for receiving l bits of data which is caused only by circuit loss.

In which:• Eelec: the energy consumption per bit in the transmitter

and receiver circuitry;

• Free space model’s amplifier energyconsumption;

• Multiple attenuation model’s amplifier energy consumption;

• a constant which relies on the applicationenvironment.

Figure 4. Network Model diagram

B. Network modelWe consider a sensor field consisting of a set of sensors

deployed randomly in a rectangular space. The algorithmassumes the following characteristics:

• Sensor nodes have similar capabilities for sensing,processing and communication.

• Sensor nodes transmit data to its immediate clusterhead in the allotted time slots or to the backup.

• All nodes are energy constrained and perform similar task.

• Network is densely populated and the sensor nodes are randomly spread in the environment.

• Nodes have been assumed homogeneous

• The sensor nodes are mobile and they are deployedrandomly in the network

• There is only one base station node that could be inside or outside the sensor network.

• The sensor nodes are unaware of its location or the position in the network.

• The batteries of the sensor nodes are non replaceable and not-rechargeable.

• Nodes are able to adjust their sending power according to their distance to the intended receiver.

• All nodes have equal energy and ability.

• Location and ID for all nodes is known for basestation.

IV. PROPOSAL- PASSIVE CLUSTERING ALGORITHM BASEDON DISTANCE AND RESIDUAL ENERGY

In this section, we present the details of the new algorithm which provides several advantages . It uses balanced energy

consumption among network nodes, minimizes the number of clusters (Clusterhead) and provides effective coverage of the network, thus it keeps longer the structure of clusters andminimize the consumed energy. As a result, the networkstability is preserved and the lifetime of the network issignificantly increased.

A. EPCDRE mechanismEPCDRE (Enhanced passive clustering algorithm based

on distance and residual energy) defines a protocol for cluster formation and election of clusterheads which assigns a weight to each node i as follows:

K(i)=En(i) ÷d0(i)En(i)= Eremaining (i)÷Einitial (4)d0(i)= d(i)÷D

Where k(i) is the threshold of elect CH, Eremaining(i) is theresidual energy of node i, d(i) is the average distance between node i with all other nodes in the same cluster and D is themaximum range of a given node.

EPCDRE is based on the following principles:

a) There are six possible states: dead, initial, ordinary, clusterhead_ready, custerhead, gateway and clusterhead-Backup.

b) Initially, all nodes are in the 'initial' state. This state does not change as long as a node does not receive a packet from another node.

c) When a node receives a packet and if the state of a sender is ClusterHead the node switches to state ordinary orgateway. Otherwise, the receiver's state switches toClusterHead_ready,

d) A node in ClusterHead_ready state will switches toClusterHead, when its coefficient K(i) is best.

e) The node ClusterHead_ready switches to stategateway when the number of ClusterHeads is greater or equal to the number of Gateways. Otherwise, the node becomes an Ordinary Node or an alternate node.

f) The node ClusterHead_ready switches toclusterhead-backup status when the number of clusterheads is greater than or equal to the number of gateways and the number of clusterheads is greater than the number of backupsand the coefficient K(i) is the second best. Otherwise, the node becomes an Ordinary Node.

g) The cluster head node selects the second best K(i)node as clusterhead-backup in case of failure of the previous one. The cluster head checks periodically the presence of his backup. In case of failure of the backup, the cluster head replays the selection process of a new backup.

h) Similarly, if the clusterhead-backup discovers theleaving of the cluterhead it switches to state ClusterHead and launch the procedure to select a backup(see Figure 1).

i) An ordinary node switches to clusterhead-backup ifits K(i) is higher. The clusterhead-backup node switches to state ordinary.

EPCDRE uses the same principles as PC for theconstruction and maintenance of clusters in wireless sensor networks. It also inherits the characteristics of the algorithmGRIDS by giving nodes with the highest level of energy to become a critical node, i.e., ClusterHead, Clusterhead-backupor GateWay.

B. state diagram EPCDRE

Sender stat = CH && Sender RE< My RE && N-CH <= N-GW && N-CH>N-BKP

Sender stat = CH && Sender RE > My RE && N-CH <= N-GW

Sender stat !=CH &&Sender RE> My RE &&N-CH>N-GW

Backup K(i)<

My K(i)>

Transmit a Packet CH Ready

My K(i)>Backup K(i)

Sender stat = CH &&Sender RE > My RE && N-CH <= N-GW && N-CH=>N -BKP CH switches to

ordinary node

N-CH<N-GW

N-CH>N-GW

Sender stat !=CH &&Sender RE>My RE

CH timeout

Sender stat!= CH

Initial

CH

Dead node

Gateway

Backup

Ordinary Node

CH timeout

Out of Power

Figure 5. State diagramme of EPCDRE

V. SIMULATIONS

In this section, we present comparison between proposed algorithms and three most important clustering protocols,PCEEC, PC and GRIDS. The simulation models used for the performance evaluation were implemented in the GloMoSim library [8]. The GloMoSim library is a scalable simulation environment for wireless network systems using the parallel discrete-event simu lation language called PARSEC [9].

A. Simulation EnvironmentThe simulation parameters used are as follows:

• the roaming space is 600X600 m square,

• The radio propagation of each node reaches up to 250 meters

• The channel capacity is 2 Mbits/second.

• The battery capacity is equal to 500 mW

• Nodes are mobiles.

• Simulations use a variable number of nodes ;distributed randomly in the roaming area;

• The random-way point model is used for node mobility

• AODV [10] is chosen as the routing protocol;

B. Simulation Results and AnalysisWe begin first by specifying the metrics that we considered

interesting to evaluate this algorithm and results obtained. Weuse three metrics for analyze and compare the simulation results: network lifetime, energy wasting and delivery ratio atbase station.

0

2000

4000

6000

8000

10000

100

200 300 400 500 600

700

800 900 1000

Number of nodes

Tota

l ene

rgy

cons

umpt

ion

PC

GRIDSPCEECEPCDRE

Figure 6. Total energy consumed during a simulation of 30 seconds.

Figure 6 show that the proposed algorithm consumes more energy for a number of nodes less than 300 nodes . By against, it consumes less energy for a greater number of nodes. Thus, we conclude that this algorithm is more suitable for large scale networks.

0%10%20%30%40%50%60%70%80%90%

[50-100] [100-150]

[150-200]

[200-250]

[250-300]

[300-350]

Time(Sec)

% o

f dea

d no

des

PCGRIDSPCEECEPCDRE

Figure 7. Percentage of dead nodes in a simulation of 300 nodes.

Figure 7 shows a comparison of dead nodes , over time, in the network using the three algorithms PC, PCEEC, GRIDSand proposed EPCDRE algorithm. The results show that the proposed algorithm retains more the energy of each node.Thus, it achieves better results in optimizing the energy consumption and prolonging the lifetime of the network.

0%10%20%30%40%50%60%70%80%90%

100%

50 100 150 200 250 300 350

Delivery ratio

Tim

e(Se

c)

PCGRIDS-PC

PCEECEPCDRE

Figure 8. Delivery ratio in a simulation of 300 nodes.

Similarly, Fig.8 shows that also the Delivery ratio is much better with EPCDRE, because EPCDRE decreases the number of dead nodes, minimize and retains more the cluster structure. Thus, the simulation results show that the Passive Clustering Algorithm Based on Distance and Residual Energy forWireless Sensor Network scheme not only provides an efficient forwarding and balances the energy consumption but alsoimproves network performance. Delivery ratio decreasesrapidly with others algorithms against EPCDRE. Figure 7 shows clearly the superiority of EPCDRE.

VI. CONCLUSION AND FUTURE WORKDue to the limitations of wireless sensor networks in terms

of energy, many algorithms for self-organization have been proposed to increase the lifetime of the network. This paper includes several contributions; first, it has considered thecritical nodes (clusterhead) and these nodes select backupnodes with more energy and less distance. The second, Nodes clusterhead and nodes backup shall periodically to shareinformation and verify the presence of each other. The third, the clusterhead and the backup clusterhead are selectedaccording to the average distance between the nodes of the cluster and the remaining energy. Simulation results show the effectiveness of the approach in reducing the amount of energy consumed by the network in comparison with three well-known protocols, passive clustering, PCEEC and GRIDS PC.

Specifically, our simulations show that:

• EPCDRE reduces communication energy by as much as compared with PC, PCEEC, and GRIDS.

• The average remaining energy of the sensor nodes in the proposed EBCRP is nearly 15% higher than the existing protocol.

• The packet delivery ratio in the proposedalgorithm is very maximal.

• EPCDRE increase the lifetime of the network

In the future, we plan to study different failure scenarios in sensor networks, introduce run-time fault-tolerance in thesystem, optimize the number of levels to efficiently consume the energy of all nodes and improve the network lifetime. We also want to extend our algorithm to heterogeneous WSNs.

REFERENCES

[1] Ian F. Akyildiz, Weilian Su, Yogesh SanKarasubramaniam, and ErdalCayirci, "A survey on Sensor Networks, "IEEE CommunicationsMagazine, vol.40, No.8, pp.102-114, August 2002.

[2] W. Heinzelman, A. Chandrakasan and H. Balakrishnan, Energy-Efficient Communication Protocol for Wireless Micro-sensor Networks, Proceedings of the 33rd Hawaii International Conference on System Sciences (HICSS '00), (2000)

[3] O. Younis and S. Fahmy, “HEED: A Hybrid, Energy -efficient,Distributed Clustering Approach for Ad Hoc Sensor Networks,” IEEE Transactions on Mobile Computing, vol. 3, no. 4, pp. 366-379, 2004.

[4] D. El-Ghanami, T.J. Kwon, A. Hafid, GRIDS: Geographically Repulsive Insomnious Distributed Sensors An efficient node selection mechanismusing Passive Clustering, IEEE WIMOB'08, France, 2008

[5] Abderrahim Maizate and Najib El Kamoun. Article:” Passive Clustering for Efficient Energy Conservation in Wireless Sensor Network”.International Journal of Advanced Computer Science and Applications,Vol. 4, No.1, 2013.

[6] T. J. Kwon and M. Gerla, Efficient flooding with passive clustering in ad hoc networks, ACM SIGCOMM Computer Communication Review, 32(1), 44–56, 2002.

[7] Heinzelman W,Chandrakasan A,Balakrishnan H. Energy-efficientCommunication Protocol for Wireless Sensor Networks[C]//Proceeding of the Hawaii International Conference System Sciences,Hawaii,January 2000.

[8] M. Gerla, L. Bajai, M. Takai, and R. Ahuja. GloMoSim: A Scalable Network Simulation Environment, Technical Report 990027, Universityof California at Berkley, 1999.

[9] R. Bagrodia, R. Meyer, M. Takai, Y. Chen, X. Zeng, J. Martin, B. Park, H. Song, Parsec: A parallel simulation environment for complexsystems, Computer, Vol. 31(10), October 1998, pp. 77-85.

[10] C. E. Perkins et al., Ad hoc on-demand distance vector (AODV) routing, [Online] Available: http://www.ietf.org/internet -drafts/draft -ietf-manet-aodv-13.TXT

[11] Loscri, V.; Morabito, G.; Marano, S. A Two -Level Hierarchy for Low-Energy Adaptive Clustering Hierarchy. In Proceedings of the 2nd IEEE Semiannual Vehicular Technology Conference, Dallas, TX, USA, 25–28September 2005; pp. 1809–1813.

[12] Fan, X.; Song, Y. Improvement on LEACH Protocol of Wireless Sensor Network. In Proceedings of International Conference on SensorTechnologies and Applications, Valencia, Spain, 14–20 October 2007, pp. 260–264.

[13] Mo Xiaoyan. Study and Design on Cluster Routing Protocols ofWireless Sensor Networks. Ph.D. Dissertation. Zhejiang University,Hangzhou, China, 2006.

[14] Heinzelman, W.B.; Chandrakasan, A.P.; Balakrishnan, H. Anapplication-specific protocol architecture for wireless microsensornetworks. IEEE Trans. Wirel. Commun. 2002, 1, 660–670.

[15] Yassein, M.B.; Al-zou’bi, A.; Khamayseh, Y.; Mardini, W.Improvement on LEACH protocol of wireless sensor network(VLEACH). Int. J. Digit. Content Technol. Appl. 2009, 3, 132–136.

[16] Ran, G.; Zhang, H.; Gong, S. Improving on LEACH protocol of wireless sensor networks using fuzzy logic. J. Inf. Comput. Sci. 2010, 7, 767–775.

[17] Abdulsalam, H.M.; Kamel, L.K. W-LEACH: Weighted Low Energy Adaptive Clustering Hierarchy aggregation algorithm for data streams in wireless sensor networks. In Proceedings of IEEE International

Conference on Data Mining Workshops (ICDMW), Sydney, Australia, 14 December 2010, pp. 1–8.

[18] Hong, J.; Kook, J.; Lee, S.; Kwon, D.; Yi, S. T -LEACH: The method of threshold-based cluster head replacement for wireless sensor networks. Inf. Syst. Front. 2009, 11, 513–521.

[19] D. J. Dechene, A. El Jardali, M. Luccini, and A. Sauer, “A Survey of Clustering Algorithms for Wireless,” Department of Electrical and

Computer Engineering The University Of Western Ontario London, Ontario, Canada