Detecting Cluster Head Attacks in Heterogeneous Wireless Sensor Networks

It is well known that clustering sensor nodes and letting those with superior resources act as cluster heads help to address the network scalability problem in wireless sensor networks (WSNs). However, the clustered architecture provides the opportunity for an adversary to attack sensor nodes within a cluster by simply compromising the cluster head. This paper investigates the attack detection issues in clustered WSNs where cluster heads are exposed to hostile environments. We propose a novel cross- layer scheme that detects abnormal cluster head behaviors. The proposed approach only makes use of the transmitted information of sensor nodes and does not require additional energy consumption at sensor nodes to monitor the cluster heads. Analytical results for detection statistics and simulations are presented to show the effectiveness of the proposed scheme.