Enhancing dependability of wireless sensor network under flooding attack: a machine learning perspective

Abstract

The wireless sensor network (WSN) is gaining paramount importance due to its application in real-time monitoring of vast geographical regions. The deployment paradigm shift is taking place from mobile computing to data science. Bridging the two technologies results in the development of dependable network in which security plays a pivotal role. This work considers the flooding attack which causes the communication failure. To detect this attack, an intrusion detection system (IDS) based on the randomised and the normalised deployment of nodes is proposed. Furthermore, machine learning techniques are implemented to enhance the dependability of network under flooding attack. The data flow is a significant parameter for governing the flooding effect on the network. It is found that machine learning models play a significant role in the prediction of the data flow. The experiments on simulated dataset underline the role of machine learning model for data flow prediction on the normalised dataset.