A novel routing with data mining in disconnected mobile ad-hoc networks

Abstract

During last two decades, Mobile Ad hoc Networks (MANETs) has been in widespread use. However, in some critical applications such as in battlefield or area sensing, due to the node mobility, limited power and sparse deployment density, MANETs may suffer network partitioning. Without a stable end to end path, these challenged scenarios could cause most existing ad hoc networks protocols to fail. In this paper, we propose a new carrier-based routing scheme with data mining method, named SARD (Self-Adaptive Routing for Disconnected MANETs). In our scheme, nodes are classified by their mobility patterns dynamically. Several Mobile Assisted-carriers (MA) are attached to each class to help data transmission. By exploiting the regularity of nodes mobility in real world, the SARD scheme improves data delivery performance in these disconnected scenarios. The simulation results show that our scheme outperforms Message Ferry approaches, and also adapts well when network scenario changes.