H.: Location management by movement prediction using mobility patterns and regional route maps

Abstract. In this paper we argue that in most of the cases the movement pattern of a mobile host repeats itself on a day-to-day basis, but for the occasional transient deviations. Taking the spatio-temporal properties of a mobile host into account, we propose a new location management scheme. The scheme achieves the near optimal routing as it bypasses the default reliance on the routes through the home agent for most of the calls made to a mobile host. It uses the mobility pattern of the mobile host to predict the cell location of that host. Transient deviations ranging from 5-30 % are tackled by tracking down a host efficiently with the help of a regional route map which is the physical route map of a small neighbourhood of the last known location of that host. The performance of the proposed scheme is evaluated with respect to varying values of callto-mobility ratio (CMR), and found to be quite good even for transient deviations ranging upto 30%