Detection of Node-Misbehavior using Overhearing and Autonomous Agents in Wireless Ad-hoc Networks

Abstract

In Wireless Ad-hoc Networks, nodes co-operate among themselves to forward data packets from a source node to a destination node. Nodes may participate in route discovery or route maintenance process but refuse to forward packets due to presence of faulty hardware or software or to save their resources, such as, battery power and bandwidth. Detection and isolation of misbehavior nodes are important issues to improve the quality of communication service and to save resources of well behaving wireless nodes. In this work, firstly, a neighbor Overhearing based Misbehavior Detection(OMD) scheme is proposed. In OMD, each node overhears the transmissions of its neighbors and calculates packet forwarding ratio of its own as well as its neighbors. Source node uses the calculated information to identify a misbehaving node. Secondly, an Autonomous Agent based Misbehavior Detection(AAMD) technique is proposed. In AAMD, past behavior of nodes in the network is used as a metric to calculate the selection probability of a node. An agent residing at a node is activated using the activation key generated by a trusted third party to verify the misbehavior of the node. The proposed schemes reduce communication overheads and identification delays to detect misbehaving nodes in wireless ad-hoc network. Simulation results are presented to evaluate the performance of the proposed OMD and AAMD schemes.