A Zone-Based Clustering Protocol for Wireless Sensor Networks

Transcription

1 A Zone-Based Clustering Protocol for Wireless Sensor Networks * S.A. SAHAAYA ARUL MARY, # GNANADURAI JASMINE BEULAH * Professor, Jayaram College of Engineering and Technology, Trichy, Tamilnadu, India # Asst. Professor, Dr. D. Y. Patil Institute of Computer Applications, Pune, Maharastra, India * samjessi@gmail.com, # jasminebsamson@gmail.com Abstract: - We study the impact of heterogeneity of nodes in the zone conception of clustered Wireless Sensor Networks. Most of the classical heterogeneous clustering protocols consider the sensing field as a whole for cluster formation. Unlike the existing protocols, the proposed protocol creates rectangular fields called zones and the number of cluster heads elected depends on the zone size and the number of nodes. The best way to extend the life span of sensor nodes is to reduce the complexity in communication costs between the sensor nodes and their cluster heads and compute optimal probability of nodes to become cluster heads in their respective zones. To improve the performance of zone-based clustered Wireless Sensor Network, we optimize some of the factors of the networks such as the number of clusters, type of cluster head elected and the total energy expended at the zones. To improve the performance of zone-based hierarchically clustered Wireless Sensor Network, we optimize certain factors of the network such as the number of clusters established, type of cluster head elected and the total energy expended at the zones. We show by simulation results that the evenly distributed energy consumption of nodes in the zones reduces the total energy dissipated in the network. Simulation results also reveal significant enhanced and favorable results on the new zone based heterogeneous setting that were not possible before. Key-Words: - Wireless sensor networks, heterogeneity, optimal probability, cluster heads, rectangular field, zone-based hierarchically clustered 1 Introduction Wireless Sensor Networks (WSNs) are extremely versatile and can be rapidly deployed to support a wide variety of applications like monitoring remote fields, surveillance of critical environments like battle field and border security where large number of sensor nodes are often unattended and work autonomously [1-4]. They contain small, multifunctional, battery-powered sensor nodes with limited energy, processing power and sensing ability. Power consumption takes place in three functional areas such as sensing, communication and data processing. A node may play a dual role of data collection and processing and also act as a data relay point for which it requires additional power. Several routing strategies have been proposed to address the design requirements of WSNs for power and resource limitations of the network nodes, possibility of packet loss and delay and timevarying quality of the wireless channel. The key challenge in the design of sensor networks is the development of energy-aware protocols that can efficiently prolong the lifetime of the network [5]. Generally, there are four main types of approaches to achieve energy efficiency, scalability and stability through data dissemination. a) Flat-network approach: All nodes are considered peers and direct transmission of data to the sink takes place. Advantages include minimal overhead to maintain the infrastructure and there is always a potential to discover multiple routes between communicating nodes for fault tolerance. b) Cluster approach: Nodes are organized as clusters and a node with higher residual energy assumes the role of cluster-head. Clustering has potential to reduce energy consumption and extend the lifetime of the network. c) Data-centric approach: Interest dissemination in the network is achieved by assigning tasks to sensor nodes and expressing queries relative to specific attributes. Its localized interactions allow it to achieve relatively high performance over un-optimized paths. d) Location-based approach: Source node queries about a phenomenon in a specific geographical area of the network where an interest may occur in a specific point in the network. The control overhead is reduced substantially as the position of the best neighbor to make correct forwarding decisions is alone sufficient. In all the four approaches, energy management is a key concern which must achieve a long lifetime while operating on limited battery reserves. As numerous applications are finding use of sensor networks, reliable information has to be delivered to the sink which is correlated both spatially and temporally [6, 7]. ISBN:

2 In a cluster-based environment, it has been proved that energy is optimally utilized when information sensed by the nodes are sent to the cluster-head (CH) [8]. Clustering can localize the route set-up process and minimize the size of the routing table in each sensor node. The cluster-head also serves as a fusion point where the data is aggregated and fused before sending it to the sink thus decreasing the redundant data [9]. Topology maintenance overhead is reduced as sensors try to connect to their CH only [10]. Optimized management strategies are implemented by the cluster heads to prolong the battery life of individual sensor nodes, network lifetime and reduce the rate of energy consumption by scheduling the activities within the cluster [11]. Various homogeneous hierarchical energy-aware routing protocols like LEACH [12], Power Efficient Gathering in Sensor Information System (PEGASIS) [13], Hybrid-Energy Efficient Distributed Clustering (HEED) [14], TEEN [15], and APTEEN [16] fall under this category. Most of the real life applications require heterogeneity rather than homogeneity with different sensing tasks. In a heterogeneous sensor network, some of the nodes are equipped with more energy than the other nodes and becomes heterogeneity in terms of node energy [17]. There are a few related works that deal with heterogeneity in the context of sensor networks. Classical approaches of heterogeneity were insufficient to fulfil the demand of balanced energy utilization [18, 19]. 1.1 Outline of the paper In this paper, we propose ZSEP-E protocol, Zone Based Stable Cluster Head Election protocol- Enhanced for clustered Wireless Sensor Networks that balances the energy levels of the nodes by efficient cluster head selection in the rectangular zones. We assume that the base station is not energy limited and that the coordinates of the base station and the dimensions of the field are known. We also assume that the nodes are uniformly distributed and that they are not mobile. The contribution of the work presents a novel network deployment model where the network is partitioned into zones. As furthest area from the base station (BS) needs more energy to transmit to the BS, high energy advanced nodes are deployed in the furthest zones of the base station. Intermediate nodes and normal nodes are deployed in the close proximity zone of the BS. Clustering technique is used locally in the zones where dense clusters are formed in the respective zones and maximum residual energy nodes are probabilistically elected as cluster heads. Deploying appropriate energy level nodes in the zones prolongs the time interval before the death of the first node which is crucial for many applications where the feedback from the sensor network must be reliable. We show by simulation that achieving this balance brings about robust performance in the network by enhancing stability period of the zones and higher average throughput than the current clustering heterogeneous zone - oblivious protocols. We also study the sensitivity of our protocol with variable size of zones along with energy heterogeneity in the sensors of the network. The rest of the paper is organized as follows. We briefly review the related work in section II with short descriptions about SEP-E and Z-SEP; motivations towards Zone-based Protocol; section III gives a description of how optimal number of clusters are constructed in a rectangular WSN; propose the protocol with network model specifications, energy consumption constraints in the stable period of the zones and deployment. In section IV simulation settings and results are analyzed. We conclude the paper in section V. 2 Related Work and Motivation Sensor nodes usually work unattended in remote geographical areas. They may be operating in a chemically or biologically contaminated field, at the bottom of oceans, in large buildings, in a battlefield beyond enemy lines. Special wireless routing protocols are essential to manage the communication between the base station and the sink. To manage high density of hundreds and thousands of sensor nodes in wireless sensor networks, clustering techniques were used. There are two broad-categories of clustered sensor networks viz. homogeneous and heterogeneous hierarchical sensor networks. In homogenous networks, all the nodes are identical in terms of energy and functionality. In heterogeneous networks, two or more different types of nodes in terms of battery energy and functionality are considered. In our paper, we consider both homogeneous and heterogeneous nodes deployed in the homogeneous and heterogeneous zones respectively. In this section, we present the literature survey of various published work. Low- Energy Adaptive Clustering Hierarchy (LEACH) [20] is a cluster-based protocol which distributes the energy consumption equally among the sensor nodes by electing cluster-heads under probabilistic rotation. However, energy drainage at the cluster-head is higher due to additional functionalities incurred such as data collection and ISBN:

3 processing. The network has to be re-clustered frequently if failure occurs at the cluster-head. Far away nodes from the sink are also susceptible to higher energy drainage due to communication costs and they tend to die quickly. Consequently, energy is not maintained uniformly between the nodes which creates energy imbalance. Distributed Energy Efficient Clustering Algorithm (DEEC) elects cluster heads based on the knowledge of the ratio between the residual energy of the nodes with that of the average energy of the entire network [21]. Additional energy is incurred to share the knowledge between the nodes. Stochastic DEEC (SDEEC) was proposed as an extension of DEEC [22]. Two-level heterogeneity is considered and energy is conserved by making non-ch nodes sleep. The disadvantage of this protocol is that during sleep state, the non-ch members are not aware of the network operations. In Distributed Energy Balance Clustering Protocol (DEBC), the result of two-level heterogeneity is extended to multi-level heterogeneity [23]. Cluster heads are elected based on the knowledge of the ratio between the remaining energy of the node and the average energy of the network. Stable Election Protocol (SEP) discusses the instability of LEACH in the presence of heterogeneity [19]. The weighted election probability of each node to become clusterhead is based on the energy of the node and does not require energy knowledge sharing. The higher energy nodes are not efficiently utilized. Z-SEP is a zone-based protocol where the sensing field is divided into zones and advanced nodes were in the furthest zone of the base station [24]. Advanced nodes were elected as cluster heads and normal nodes directly communicated with the base station. But conventional protocols like Direct Transmission (DT) do not assure a balanced and uniform use of the sensor energy. SEP-E, a modified algorithm of SEP considers three-tier nodes by introducing intermediate nodes which serves as a bridge between normal nodes and advanced nodes [18]. SEP-E does not provide an ideal solution for energy dissipation, since normal nodes, intermediate nodes and advanced nodes are randomly deployed. If most of the normal nodes and intermediate nodes are deployed far away from the base station, more energy is consumed for transmitting data which results in the decrease of throughput and network lifetime. To give a better solution for the above problem, we propose zone based protocol where the network is partitioned into zones. 2.1 Energy Heterogeneity of SEP-E In this section, we briefly discuss the idea behind SEP-E and its improvement over the SEP. Aderohunmu et al. proposed SEP-E and improved the stable region of WSN by considering three-level energy heterogeneity [18]. The new epoch accommodated the extra energy factor introduced over SEP and is equal to 1/p opt (1+mα+bµ). A weight equal to the energy of the node divided by the relative energy of other nodes was used as an election probability for the cluster heads in SEP-E. The probability for a node to be elected as CH and total initial energy of the system is given below: (1) where P nrm,p int and P adv are the weighted probabilities of normal nodes, intermediate nodes and advanced nodes respectively. b and m are the proportions of the intermediate and advanced nodes with α and µ times more energy than the normal nodes respectively. E total is the total initial energy of the system. This protocol considers the whole sensing field to elect cluster heads and the cluster heads may be located near the edges of the network and as a result the nodes have to bridge long distances to reach the base station. This leads to high energy consumption and energy imbalance since nodes have to transmit over long distances. 2.2 Zone Based Two-Level Energy Heterogeneity of Z-SEP In this section, we precisely discuss Z-SEP for two-level heterogeneity proposed by Faisal et al. [24]. Z-SEP divides the sensing field into three zones as Zone0, Head Zone1 and Head Zone2. Normal nodes were deployed in Zone0 and they transmit data directly to the base station using Direct Transmission technique. Advanced nodes were randomly deployed in Head Zone1 and Head Zone2 and clusters were formed, cluster heads were elected to send the sensed data to the base station. In this protocol, normal nodes communicate directly to the base station using Direct Transmission (DT) technique and they tend to die quickly. (2) (3) (4) ISBN:

4 2.3 Motivation In SEP-E, the normal nodes, intermediate nodes and the advanced nodes are randomly deployed. The cluster-head election is based on the election probabilities weighted by the energy of a node relative to other nodes in the network. The normal nodes elected as cluster heads furthest from the base station are the ones to die out first since they expend large communication energy. This results in a large increase in energy dissipation as communication is much higher than computation. By analyzing the advantages and disadvantages of heterogeneous routing protocols, we propose our protocol to minimize-energy dissipation in sensor networks. In this work, we consider the sensor network where: The base station is fixed and located at the centre of the sensor field with its dimensions known. The entire sensor network is partitioned into two homogenous zones (furthest zones with BS) and one heterogeneous zone (near-by zone with BS). In the homogenous zone, the advanced nodes are deployed to form clusters and cluster heads are elected. In the heterogeneous zone, the normal and intermediate nodes form clusters and the intermediate nodes are elected as cluster heads. All the nodes in the network are heterogeneous and energy-constrained. Key features of our protocol are: Zone localized coordination and control for cluster setup and operations. Randomized election and rotation of high energy nodes as cluster-heads in the clusters of the respective zones. Data aggregation and fusion at the cluster heads for larger energy gains. 3 Optimal Construction of Clusters in Rectangular Fields of WSN The optimal probability of a node to become a Cluster Head CH is a function of spatial density when energy of the nodes is uniformly distributed over the sensor field. Optimal clustering depends on the radio energy dissipation model used and the optimal clustering is attained when energy consumption is well distributed over all the sensors and the total energy consumption is minimum. We follow the radio energy dissipation model of the previous works given by Aderohunmu et al. [18]; Smaragdakis et al. [19]; Heinzelman et al. [20]. Fig.1 shows the radio energy dissipation model. Fig.1 Radio Energy Dissipation Model According to the model, the energy expended by the radio to overcome the free space (fs) or multipath loss (mp) and to transmit k bits of data over a distance d in the acceptable Signal-to-Noise Ratio is given by (5), where E elec =50nJ/bit energy needed to run the transmitter or receiver circuit, d 0 is the distance threshold and is calculated by equating the two expressions d=d 0, we have do= d o = To receive k bits of data, the radio expends, E Rx = ke elec. We assume an area A= a b square meters, and n is the number of nodes that are uniformly distributed over that area and the BS is located in the centre of the field and the distance of any node to its BS is d 0. During a round, the energy dissipated by a CH is given by the equation (6), (6) Where k is the number of clusters, E DA is the data aggregation cost of a bit per signal and d tobs is the average distance between a CH to the BS and the energy consumed between them is given by the equation (7) as in Saadi et al. [25], (7) The non-cluster head, NCH dissipates energy equal where d toch is the average distance between a cluster member and its cluster head and it is given by the equation (9), (8) (9) (5) ISBN:

5 The total energy expended by the cluster per round, E clustertot is given by the equation (10), (10) The area A is partitioned into three zones as follows and shown in the Fig.2 Zone Based Heterogeneous Setting. The total energy expended by the network is given by the equation (11), The optimal number of clusters can be found by differentiating E tot with respect to k and equating to zero and is given by the equation (12), (11) = (12) The optimal probability of a node, p opt to become a cluster head is computed by the equation (13), (13) Heinzelman et al. showed that if the clusters are not constructed in an optimal way, the total energy expended by the sensor network would increase exponentially either when the number of clusters is greater or less than the optimal value [20]. 3.1 Zone Based Clustering Algorithm The performance and lifetime of WSNs is highly influenced by the clustering scheme. But the efficiency of the network is drastically affected by the early death of the sensor nodes which are furthest away from the base station. To address this issue we propose a new Zone-Based protocol by partitioning the field, to deploy the nodes efficiently in the sensing field, organize them into clusters in an effective way and elect suitable cluster heads in the zone. Partition of the network into zones improves the coverage and connectivity of the cluster nodes which are far away from the base station. 3.2 Network Model and Specifications We have made some assumptions about the underlying network model of area A= X Y sq.mts. where X=a, Y= b 1 +b 2 +b 3, where b 1 =b 3 and b 1 +b 2 +b 3 =a and the sensor nodes deployed. Fig.2. Zone Based Heterogeneous Setting Each zone is a geographical division of the sensing field and considered separately and appropriate energy-level sensor nodes are deployed depending on the distance and orientation from the base station. Let m proportion of total number of nodes n are equipped with α times more energy than the normal nodes and it is referred as advanced nodes. Let b proportion of total number of nodes n are equipped with µ times more energy than the normal nodes and it is referred as intermediate nodes. 1. Zone1 = a b 1, lying between 0 Y Y 1 deployed with static homogenous energy-rich advanced nodes where. 2. Zone2 = a b 2, lying between Y 1 <Y Y 2 deployed with b proportion of static intermediate nodes and (1-m-b) n normal nodes where 3. Zone3 = a b 3, lying between Y 2 <Y Y 3 deployed with static homogenous energy-rich advanced nodes where. The main design objective to deploy identical advanced nodes at Zone1 and Zone3, described as the furthest zone from the base station in our paper, is to guarantee certain network lifetime and to ensure that all the nodes expire at about the same time. In the homogenous zone, all the nodes are capable of acting as a cluster head. The failure of few nodes does not seriously affect the working of the scheme and thus the zone is robust to node failures. ISBN:

6 3.3 Zone Based Clustering Scheme In the proposed protocol, we assume that the base station, BS is static and located at the centre of the field. A hierarchical clustering protocol is employed for all the zones where the nodes are self-organized into clusters and local base station or cluster heads, CH are elected. At a given time, the intermediate nodes and the advanced nodes in respective zones elect themselves as cluster heads with a certain probability and broadcast their status to other nodes in their respective zones. Each sensor node determines their cluster head based on the minimum communication energy to the CH. The cluster head election is based on the randomized weighted election probabilities according to the residual energy of the node in the cluster of their respective zones. The cluster head create a schedule for communication for the nodes in its cluster once all the clusters are organized. The cluster members sense the data and transmit them to their CH. To minimize the energy dissipated by the non cluster heads, the radio components are turned off at all times except during its transmission time. Once, when all the data is available at the CH, it aggregates the data and sends them to the BS. Further to reduce energy dissipation and enhance network lifetime, CH adopts local data fusion to compress the amount of data being sent from the clusters to the base station. To efficiently use the battery of every single sensor, the cluster heads are re-established at every round. The operation of the protocol is broken down into rounds with each round consisting of cluster set-up and the steady- state phase. The clusters are formed in the cluster set-up phase and data transmission from the cluster head to the base station takes place in the steady-state phase. 3.4 Energy Consumption constraints in the Stable period of the Zones The constraint of well balanced energy consumption has to be maintained in order to prolong the stable region. A fairness constraint in energy consumption is achieved by having the advanced nodes and intermediate nodes to become cluster heads at every round. The heterogeneous setting in Zone2 (with intermediate nodes and normal nodes) and the homogeneous setting in Zone1 and Zone3(with advanced nodes) will have more effect on the spatial density of the network and so the apriori setting of p opt, the optimal probability of a node to become cluster head in the zone changes. The optimal probability of a node to become a cluster head in each Zone is computed as follows Let p opt1 be the optimal probability of an advanced node to become a cluster head in Zone1 and Zone3 and it can be calculated as follows, Where k opt1 is the optimal number of clusters constructed in Zone1 and Zone3. Let p opt2 be the optimal probability of an intermediate node to become a cluster head in Zone2 and it can be calculated as follows, where k opt2 is the optimal number of clusters constructed in Zone2 and n is the total number of nodes in the network. The total energy of the system in the zones also changes. Suppose that E in be the initial energy of each of the normal node then the energy of the advanced nodes and intermediate nodes will be and respectively. The total initial energy of the system in Zone1 and Zone3, E tot1 is given by (16), The total initial energy of the system in Zone2, is given by, Hence the total initial energy of the system given by, (14) (15) (17) (16) where m and b are the proportion of the advanced (18) nodes and intermediate nodes to the total number of nodes n. The following conditions satisfy, for the stable region of the zones: 1. At Zone1 and Zone3, the advanced nodes become cluster head exactly is ISBN:

7 times every rounds. This period is defined as Zone1, Zone3 epoch. 2. At Zone2, the intermediate nodes become cluster head exactly 1+ µ times every rounds. We define this period as Zone2 epoch. 3. Average number of clusters in the network will be, (19) To guarantee a well distributed energy consumption constraints in the stable period of the zones, we assign a weight to the optimal probability p opt1 and p opt2. This weight must be equal to the initial energy of each node divided by the initial energy of the normal node. We define and as the weighted election probabilities for intermediate nodes and advanced nodes in the respective zones. There are nodes with energy equal to the initial energy of a normal node. We solve the above condition mathematically by translating it into a probability problem. Let and be the weighted probabilities of intermediate nodes and advanced nodes becoming Cluster Heads in their respective Zones, then we have: where is the set of advanced nodes that has not become cluster heads in the past rounds of the Zone1 and Zone3 epoch. is the threshold applied to a population of nm nodes. The average total number of cluster heads per round in Zone1 and Zone3 will be The average total number of cluster heads per round in Zone2 will be (24) From Eq. (23), (24), we find that the total average number of cluster heads per round for the whole system will be: (25) Because of the zone based energy heterogeneity and the type of cluster head election, the energy dissipation in the network is better controlled. 3.5 Deployment The protocol proposed for wireless sensor networks is application specific and it is hard-wired to perform specific tasks efficiently at low cost. There is always a need to re-program a group of sensors when there is a catastrophic failure of nodes. A reliable transport protocol as the one proposed in [27] could be used to re-program or re-task such sensors. Evaluating the overhead of such a deployment is a subject of our continuing work. (20) We define a threshold for the Cluster Head election process in each round to guarantee that sensor nodes must become cluster heads. The threshold and of intermediate and advanced nodes respectively is proposed by Aderohunmu et al. as follows: (21) where is the set of intermediate nodes that has not become cluster heads in the past rounds of Zone2 epoch. is the threshold applied to a population of nb nodes. Similarly, for the advanced node, we have 4 Simulation Results We simulate a clustered wireless sensor network with dimensions 100m 100m and partition the field into three zones as per our proposed protocol in MATLAB. Table I shows the various simulation parameters used in the protocol. TABLE I SIMULATION PARAMETERS Parameters Values Initial Energy of Normal nodes E in 0.5 J Initial Energy of Intermediate nodes E im (1+ µ) E in Initial Energy of Advanced nodes E ia (1+ α) E in Data Aggregation(E DA ) 5 nj/bit/signal Transmitting/Receiving Energy(E elec ) 50nJ/bit Short Distance Amplification Energy(E fs ) 10 pj/bit/m 2 Long Distance Amplification Energy(E amp ) pj/bit/m 4 Probability P opt1 (advanced node as CH) 0.2 Probability P opt2 (intermediate node as CH) (22) ISBN:

8 The total population (n) of the sensors randomly deployed in all the zones are 100. Advanced nodes with α =3 times more energy than normal nodes are deployed in Zone1 and Zone3 equally. In Zone2, intermediate nodes with µ = 1.5 times more energy than the normal nodes are distributed along with normal nodes.in order to show the sensitivity of our algorithm on different size of the zones for network lifetime and throughput, we simulate the network with different zone dimensions. Case 1: Network Lifetime and Throughput in the presence of high energy heterogeneity with zone dimensions in 100m 100m network setting. The total population (n) of the sensors randomly deployed in all the zones are 100,out of which there are 20% of advanced nodes i.e. m=0.2. Hence, the zone dimensions will be Zone1 lying between 0<Y 1 10, Zone2 lying between 10<Y 2 90 and Zone3 lying between 90<Y as shown in Fig.3 (a). Fig. 3(b) Fig.3(c) Fig. 3 Performance of ZSEP-E with m=0.2, b=0.3, α=3 and µ=1.5, SEP-E with m=0.2, b=0.3, α=3 and µ=1.5 and Z-SEP with m=0.3, b=0, α=3.5 and µ=0, E total = 102.5J. Zone area Y 1=10, Y 2=90 and Y 3=100 (b) Alive nodes per round (c) Throughput Fig.3 (a) Node distribution with Y 1=10, Y 2=90 and Y 3=100 In order to have a fair comparison with Z-SEP and SEP-E, we maintain the same zone based heterogeneous setting of nodes and also the same energy levels in all the zones as in our protocol, so that total initial energy of the system is the same in Zone-based, Z-SEP and SEP-E. We compare the performance of our protocol with m=0.2, b=0.3, α=3 and µ=1.5, SEP-E with m=0.2, b=0.3, α=3 and µ=1.5 and Z-SEP with m=0.3, b=0, α=3.5 and µ=0, E total = 102.5J in the presence of high energy heterogeneity as shown in Fig. 3(b), (c). Other parameters are shown in Table 1. We compare the performance of our protocol with m=0.1, b=0.2, α=2 and µ=1, SEP-E with m=0.1, b=0.2, α=2 and µ=1 and Z-SEP with m=0.3, b=0, α=1.3 and µ=0, E total = 70J in the presence of low energy heterogeneity as shown in Fig. 3(b), (c). Case 2: Network Lifetime and Throughput in the presence of low energy heterogeneity with zone dimensions in 100m 100m network setting. The total population (n) of the sensors randomly deployed in all the zones are 100,out of which there are 10% of advanced nodes i.e. m=0.1. Hence, the zone dimensions will be Zone1 lying between 0<Y 1 5, Zone2 lying between 5<Y 2 95 and Zone3 lying between 95<Y as shown in Fig. 4(a). Fig.4 (a) Node distribution with Y 1=5, Y 2=95 and Y 3=100 ISBN:

9 Fig.3 and Fig.4 in case 1 and case 2 respectively shows the detailed behaviour of SEP-E and Z-SEP for different distributions of heterogeneity and zone dimensions. Fig.3 (a) illustrates that the sensing field becomes very sparse quickly since the nodes tend to die very quickly in both SEP-E and Z-SEP. When a significant number of nodes are dead, then the average number of cluster heads per round is less than one. It means that in most of the rounds there are no cluster heads. Our protocol takes the advantage of deploying the extra initial energy of the advanced nodes in far-away areas of the base station. The stable region is thus far extended when compared to SEP-E and Z-SEP. The protocol also takes the advantage of the zone setting and random weighted probabilistic election of cluster heads in less spatial density area. Fig. 3(b) and 4(b) shows the sensitivity of our protocol in terms of the length of the stability period by varying m and α. The performance does not depend only on m but depends on m and a. The advanced nodes are uniformly distributed in the far-away zones of the sensing field and when they elect themselves as cluster heads, their extra energy is judiciously consumed. The energy dissipation aspect also shows the superiority of ZSEP-E compared to Z-SEP and SEP- E. Fig.5 shows the average rate of energy dissipation pattern. Fig. 5 Rate of Energy Dissipation m=0.2, b=0.3, α=3 and µ=1.5 ZSEP-E s energy dissipation slope for normal nodes is much flatter when compared to SEP-E and Z-SEP showing a prolonged stability round achieving 2010 rounds. The energy dissipation of intermediate and advanced nodes are more violent in ZSEP-E than SEP-E and Z-SEP since these nodes serve as cluster heads more often than the normal nodes, which reduces the energy consumption of normal nodes. This clearly shows that the new zone-based design achieves better utilization of energy with uniform energy drainage. Fig. 4(a) Fig. 4(b) Fig. 4 Performance of Zone based protocol with m=0.1, b=0.2, α=2 and µ=1, SEP-E with m=0.1, b=0.2, α=2 and µ=1 and Z-SEP with m=0.3, b=0, α=1.3 and µ=0, E total = 70J Zone area Y 1=5, Y 2=95 and Y 3=100 (b) Alive nodes per round (c) Throughput 5 Conclusion We proposed ZSEP-E, a Zone-based protocol by deploying specific type of energy-level nodes from the location of the BS with cluster formation within rectangular zones and the cluster head election is based on the probabilistic election of weighted residual energy of the nodes relative to other nodes in their respective zones. Dense clusters are formed locally in the zones. The key idea proposed here is that the network field is divided into rectangular zones and the number of cluster heads elected is determined in proportion to the number of nodes and the size of the zone. Furthermore unlike Z-SEP, our protocol increases the stability region by intelligent election of advanced nodes as cluster heads in the furthest zone of the base station and intermediate zones in the nearby zone of BS thus avoiding the direct transmission technique of Z- SEP. Simulation results show that fair distribution of sensor nodes achieves balanced consumption of energy thereby enhancing network lifetime and stability of WSNs. A potential approach in our future work would be to further partition the zones ISBN:

10 into regions and adopt multi-hierarchical clustering technique to enhance network lifetime. References: [1] T Bookareva, W Hu, S Kanhere, B Ristic, N Gordan, T Bessell, M Rutten, S Jha, Wireless Sensor Networks for battlefield surveillance, [2] D Dudek, C Haas, A Kuntz, M Zitterbart, D Kruger, P Rothenpieler, D Pfisterer, S Fischer, A wireless sensor network for border surveillance, ACM Conference on Embedded Networked Sensor Systems : , [3] J K Hart, K Martinez, Environmental Sensor Networks: A revolution in the earth system science: Earth-Science Reviews 78: [4] M Quaritsch, K Kruggl, K Wischonnig- Strucl, D Bhattacharya, S Shah, B Rinner, Networked UAVs as aerial sensor network for disaster management applications: E & I Elektrotechnik and Informations technik, 127(3): 56-63, [5] Akyildiz, W Su, Y Sankara Subramanian, E Cayirci, A Survey on Sensor Networks, IEEE Communications Magazine, 40(4): , [6] T Shu, M Krunz, Coverage-time optimization for clustered wireless sensor networks: a power-balancing approach IEEE/ACM Transactions Networking 18(1): , [7] M C Vuran, I F Akyildiz, Spatial correlation-based collaborative medium access control in wireless sensor networks. IEEE/ACM Trans Networking 14(2): , [8] K Akkaya, M Younis, A survey on routing protocols for wireless sensor networks, Elsevier Journal of Ad Hoc Networks 3(3): , [9] K Dasgupta, K Kalpakis, P Namjoshi, An efficient clustering-based heuristic for data gathering and aggregation in sensor networks, IEEE Wireless Communications and Network Conference, [10] Y T Hou, Y Shi, H D Sherali, On energy provisioning and relay node placement for Wireless Sensor Networks, IEEE Transactions on Wireless Communications 4(5): [11] M Younis, M Youssef and K Arisha, Energy-aware management in clusterbased sensor networks. Computer Networks 43(5): [12] W Heinzelman, A Chandrakasan, H Balakrishnan, An application-specific protocol architecture for wireless microsensor networks, IEEE Transactions on Wireless Communication, 1(4): , [13] S Lindsey, C Raghavendra, Pegasis: Powerefficient gathering in sensor information systems, in Aerospace conference proceedings, 3 (3):11-24, [14] O Younis, S Fahmy, Heed: a hybrid, energy-efficient, distributed clustering approach for adhoc sensor networks, IEEE Transactions on Mobile Computing, 3(4): , [15] A Manjeshwar, D P Agrawal, TEEN: a protocol for enhanced efficiency in wireless sensor networks, International Workshop on Parallel and Distributed Computing Issues in Wireless Networks and Mobile Computing, [16] A Manjeshwar, D P Agrawal, APTEEN: a hybrid protocol for efficient routing and comprehensive information retrieval in wireless sensor networks, Parallel and Distributed Processing Symposium: , [17] V Mhatre, C Rosenberg, Homogeneous Vs heterogeneous clustered sensor networks: A comparative study, IEEE International Conference on Communications, [18] F A Aderohunmu, J D Deng, SEP-E: An Enhanced Stable Election Protocol (SEP) for Clustered Heterogeneous WSN, Discussion Paper Series, No. 2009/07. Department of Information Science, University of Otago (ISSN: X), [19] G Smaragdakis, I Matta, A Bestavro, SEP: A stable election protocol for clustered heterogeneous wireless sensor networks, International Workshop on SANPA: , [20]W Heinzelman, A Chandrakasan, H Balakrishnan, Energy Efficient Communication Protocol for Wireless Microsensor Networks, Hawaii International Conference on System Sciences, ISBN:

11 [21] L Qing, Q Zhu, M Wang, Design of a distributed energy-efficient clustering algorithm for heterogeneous wireless sensor networks. Computer Communications: , [22] B Elbhiri, R Saadane, B Aboutajdine, Stochastic Distributed Energy-Efficient Clustering for heterogeneous wireless sensor networks, ICGST-CNIR Journal, 9(2): [23] C Duan, H Fan, A Distributed Energy Balance Clustering Protocol for Heterogeneous Wireless Sensor Networks, International Conference on Wireless Communications, Networking and Mobile Computing,: , [24] S Faisal, N Javaid, A Javaid, M A Khan, S H Bouk, Z A Khan, Z-SEP: Zonal Stable Election Protocol for Wireless Sensor Networks, Journal of Basic and Applied Scientific Research (JBASR), [25] M Saadi, M. Lahcen Hasnaou, Abderrahim BeniHssane, Mohamed Laghdir, Computation of the optimal probability of becoming a cluster head in hierarchical clustered WSNs International Journal of Computer Applications ( ): 75(1), [26] Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specification, IEEE Std , [27] C.-Y. Wan, A. T. Campbell, and L. Krishnamurthy. PSFQ: a reliable transport protocol for wireless sensor networks. In 2nd IEEE Workshop on Applications and Services in Wireless Networks, pp. 1 11, July ISBN: