Hybrid Anomaly Detection by Using Clustering for Wireless Sensor Network

Performance of wireless sensor network are highly prone to network anomalies particularly to misdirection attacks and blackhole attacks. Therefor intrusion detection system has a key role in WSN and it's essential in security application. However the identification of active attacks is cumbersome in many cases particularly for remote sensing applications. This paper proposes hybrid anomaly detection method for misdirection and blackhole attacks by employing K-medoid customized clustering technique. A synthetic dataset was established by defining network parameters and threshold values were obtained to detect the anomalies. Experimental work was performed on network simulator (NS-2) and R studio. The proposed algorithm successfully detect the hybrid anomalies with high accuracy. This work is suitable for hybrid anomaly detection including misdirection and blackhole attacks in wireless environment.