Predictive Routing in Mobile Ad-Hoc Networks

Predicting future state of dynamic networks has been a growing area of research in recent years. Mobile ad-hoc networks in particular are subject to frequent disconnects and topology changes as a consequence of user mobility. However, in many kinds of mobile networks, e.g. vehicular networks, mobility is not strictly random, and some prediction of node mobility is possible, which enables estimation of route stability. However, predictive techniques rely on accurate localization, which is not always available. In this paper, we examine the effect localization error as commonly encountered in practice has on a simple mobility prediction method.