RUHEED- Rotated Unequal Clustering Algorithm For Wireless Sensor Networks Nueraili Aierken, Roberto Gagliardi, Leonardo Mostarda, Zaib Ullah Computer Science Division University of Camerino, Camerino, Italy Email: {nueraili.aierken,roberto.gagliardi,leonardo.mostarda,zaibullah.zaibullah}@unicam.it Abstract—Prolonging the network lifetime, scalability and balancing are very important requirements when implementing a wireless sensor network (WSN). Clustering is a technique that has been widely applied for achieving these goals. However, there exists the energy hole problem which causes an unbal- anced energy consumption in equally formed clusters. More specifically, nodes near the base station (BS) die very quickly since they, not only transmit their own data, but also forwards the rest of the network data. In this article, we propose a rotated unequal clustering protocol RUHEED in order to mitigate the energy hole problem. Our experiments show that RUHEED improves the network lifetime when compared to other clustering protocols. Keywords-Wireless sensor networks; clustering protocols; energy efficiency; I. I NTRODUCTION A wireless sensor network (WSN) is composed of spa- tially distributed sensor nodes with limited memory, pro- cessing, communication and energy capabilities [2], [3], [4] and a Base Station (BS). Sensor nodes sense the environ- ment, process the data and transmit it to a nearby node until they reach the BS [1]. The role of a base station is to collect all data received from the various sensors, analyse them and ultimately make decisions. WSNs can be classified as continuous or event-based [5], [6]. In the first case readings are collected and reported periodically irrespective of the changes involved. In the second case a report is sent only when a specific event happens. Depending on the application, sensor node deployment can be either deterministic or random, stationary or mobile, homogeneous or heterogeneous [7]. The energy efficiency challenge is very important when building WSNs since nodes are often equipped with a non- renewable source of energy. Thus WSNs have a bounded lifetime [8], [9]. Clustering is one of the solutions pro- posed by the researchers. Clustering organizes the WSN into sets (clusters). A sensor from each set is elected as cluster head (CH). A CH coordinates and aggregates data of nodes within its cluster (intra-cluster communication). CHs communicate with each other and/or with an external base station (BS) (inter-cluster communication) on behalf of their nodes. When equal size clustering [12], [11] is used, the energy hole problem can take place [10] where nodes close to the base station die earlier. In fact, they are burdened with heavy relay traffic from the rest of the network (intra-cluster communication) in addition to their own intra-cluster traffic (inter-cluster communication). A quit common solution to the energy hole problem is unequal clustering [10], [14], [17], [18] where clusters far away from the BS will have a larger size when compared to clusters closer to the BS. By creating unequal sized clusters, the amount of intra-cluster traffic is considerably reduced for the CHs nearer to the BS. In this paper, we present Rotated Unequal Clustering Protocol (RUHEED) that enhances the Unequal Clustering Algorithm (UHEED) [10] with the addition of rotation for cluster head election. More precisely, RUHEED does not always perform a full leader election protocol but CHs are rotated among the member nodes of the same cluster. This rotation is based on the highest residual energy. We com- pared RHEED with LEACH, HEED and UHEED. RUHEED outperforms all the mentioned clustering protocols when the first-node dies and half-nodes die network lifetime measures are considered. Section II details RUHEED and Section III discusses the network model, the simulation parameters and the results. Section IV compares the presented work with existing related works in this area. Finally, Section V provides a conclusion and outlines future work. II. RUHEED- ROTATED UNEQUAL CLUSTERING ALGORITHM The RUHEED protocol is an enhancement of the UHEED [10] that is based on HEED. Because of this evolutionary order in the following we first sketch HEED, afterwards we discuss UHEED and we conclude with the enhancement offered by RUHEED. A. HEED HEED is a clustering protocol that periodically selects cluster heads. The selection is based on different parameters such as residual energy, node degree, node proximity to its neighbors and so on. A periodical and careful change of cluster heads can balance load among clusters prolonging the network lifetime. HEED election is composed of three main phases that are initializing, iterative and final. The initializing phase assigns to each node the probability of becoming a tentative cluster head. This is done according to the following formula ( see [12] for details about the formula):