Anomaly Detection Based on Data-Mining for Routing Attacks in Wireless Sensor Networks

With the increasing deployment of wireless sensor devices and networks, security becomes a critical challenge for sensor networks. In this paper, a scheme using association algorithm and clustering algorithm is proposed for routing anomaly detection in wireless sensor networks. The scheme uses the Apriori algorithm to extract traffic patterns from both routing table and network traffic packets and subsequently the K-means cluster algorithm adaptively generates a detection model. Through the combination of these two algorithms, routing attacks can be detected effectively and automatically. The main advantage of the proposed approach is that it is able to detect new attacks that have not previously been seen Moreover, the proposed detection scheme is based on no priori knowledge and then can be applied to a wide range of different sensor networks for a variety of routing attacks.