A new weight based self-organizing approach to clustering in mobile ad hoc networks

Abstract

Mobile Ad Hoc Networks (MANET) is a collection of several wireless hosts, which form a network, without any infrastructure requirements. These are temporary networks, which cease to exist when no longer needed. In MANETs any node can act as a router, to provide connectivity to other nodes, within its transmission range. Several nodes in a Mobile Ad Hoc Network get connected via Shortest Paths to ensure minimal cost of transmission. In the process of computing the shortest path, certain critical points called articulation points are formed. Articulation points are those, when removed from the network, detaches a sub tree of nodes from the parent network, thereby disrupting the network connection. Therefore, if the node at the articulation point fails, other nodes, dependant on that node, are removed from the network. We propose an efficient method to identify and avoid formation of such articulation points, during the formation of an Ad Hoc Network. Our algorithm helps in finding alternate paths for connection, and hence improves the reliability and longevity of the network.