An anomaly detection technique-based intrusion detection system for wireless sensor network

Abstract

In Wireless Sensor Networks (WSN), it is challenging to detect the compromised node. A compromised node is a kind of node which doesn't have features like another normal node, and it can be an intruder or harmful for the network. In the homogeneous WSN, nodes have similar behaviour; this is the main reason that detecting compromised nodes is challenging. We have applied the anomaly technique to identify compromised nodes. We have chosen four specific parameters which decide the behaviour of nodes; we have selected some Intrusion Detection System (IDS) operator and trained them. Selection will be made by internal congestion within a node. The IDS operators has been chosen in such a manner that every node is in the radio range of at least one IDS operator. Learning of IDS operator is based on behaviour's parameters of nodes. After detection of an intruder, the IDS will notify the sink node.