Abstract
In recent years, a new paradigm of sensor networks, mobile sensor networks (MSNs), are emerging. They inherit all the properties from static wireless sensor networks (WSNs) and meanwhile have their own uniqueness, node mobility. Cluster-based architecture in MSNs have the advantage of providing scalable and resource efficient solutions. Thus, finding an optimal way to generate clusters is an important topic in MSNs. In this paper, we present a prediction-based clustering algorithm for MSNs. Our proposed algorithm uses adequate criterion for cluster-head election in conjunction with heuristic predictors to generate steady and balanced clusters. The results of the simulation experiments clearly show the benefit of the proposed algorithm in terms of network life time.