A comparative study of mobility prediction schemes for GLS location service

The performance of geographic routing suffers from mobility-induced location errors. Location errors can also occur due to infrequent and/or lost updates to location servers, especially when the nodes are highly mobile. In GLS (grid location service) the location update frequency to the higher order location servers is very low. A query to a location server fails when a node moves far away from its previous location, rendering the cached location in the location servers invalid. In this work, we present a detailed study on the impact of node velocity on query failure rates and failure reasons in GLS, across different mobility models. Correct and efficient mobility prediction by the location servers themselves can improve the query success rate in the grid/GLS framework. We present a comparative study to investigate the performance of various prediction schemes with GLS over a rich set of mobility models.