Abstract
|
In a real mobile ad-hoc network, a node usually moves to a pre-specified destination, and often exhibit non-random motion behavior. In such mobility patterns, the future motion behavior of a mobile node is correlated with its past and current mobility characteristics. In this paper, we propose an adaptive link lifetime prediction method. In our method, the prediction is made by exploiting the correlation of mobility parameters over time. Using curve fitting method, the proposed method learns how to predict the lifetime of a link between two adjacent mobile nodes based on their current related velocity and their distance history. Therefore, there is no need to extra information, e.g., signal strength, radio condition. Also in our algorithm, we do not impose any heavy calculation to the mobile nodes. Simulation results show the proposed algorithm predicts a link lifetime efficiently.
|