Novel intrusion detection methods for security of wireless sensor network

Abstract

Wireless Sensor Network (WSN) has to face many threats as it consists of sensor nodes and needs installation in the open area. Intrusion Detection System (IDS) is an essential safety method of handling vulnerabilities and threats for WSN. This study is a relative assessment of the best performed IDS methods of WSNs, the analysis of this method is technically represented in detail. Threats to WSN are categorized into the criteria. Customized dataset is prepared by KDD dataset with five stages to normalize it. Normal class has four types of attacks which are much related attributes and used for classification routine. This study is applied to methods (e.g. CfsSubsetEval and BestFirst) for selection of attributes procedure to remove irrelevant attributes. Experimental work reports the algorithm which provides high detection rate. Finally, in conclusion which has satisfactory statements and rules for future research works to implement IDS in WSNs. Many recommendations are mentioned as well as future directions regarding this study.

Keywords: Wireless Sensor Network, Anomaly Detection, Intrusion Detection System, Classification, KDD Dataset