Outlier detection to Secure Wireless Sensor Networks Based on iForest

Abstract

Wireless sensor networks (WSN) make use of low-cost and multifunctional nodes with limited at short distances utilizing wireless connectivity. It's an open media platform and it's based on application-driven technology. In addition to military applications, it can be used for environmental monitoring, the health sector, and to respond to emergencies as well. WSN security had attracted a great deal of attention due to its nature and application scenario. Construction of nodes and distributed network infrastructure is prone to a variety of attacks. To ensure its functionality of WSNs, security mechanisms are essential. In the WSN, one of the concerns is provisioning of the security with less powerful nodes and small network. The Isolation Forest algorithm is used in this study to detect the outlier nodes in WSN to secure the network. The research was carried out with a hundred nodes WSNs for temperature data. The result gave a promising result to implement in hardware structure in future.