Stable clustering through mobility prediction for large-scale multihop intelligent ad hoc networks

In this paper we present a framework for dynamically organizing mobile nodes (MNs) in large-scale mobile ad hoc networks (MANETs), with the eventual aim to support quality of service (QoS). Our dynamic, distributed clustering approach is based on intelligent mobility prediction that enables each MN to anticipate the availability of its neighbors. We present a scalable way to predict the mobility, and thus availability, of MNs, achieved with the introduction of geographically-oriented virtual clusters. We name the proposed model as the (p, t, d)-clustering model that facilitates the formation of stable clusters. Simulation results demonstrate the performance advantages of our approach