A New Stable Clustering Scheme for Highly Mobile Ad Hoc Networks

This paper addresses the clustering problem for highly mobile ad hoc networks. In the proposed scheme, Doppler shifts associated with received signal are used to estimate the relative speed between cluster head and cluster members. With the estimated speed, a node can predict its stay time in every nearby cluster. In the initial clustering stage, a node joins a cluster that can provide it with the longest stay time in order to reduce the number of re-affiliations. In the cluster maintaining stage, strategies are designed to help node cope with connection loss caused by channel fading and node mobility. Simulation results show that the proposed clustering scheme can reduce the number of re-affiliations and the average disconnection time compared with previous schemes.