Method for Detecting Manipulated Compilation of Sensing Reports in Wireless Sensor Networks

In cluster-based wireless sensor networks (WSNs), a few sensor nodes, including cluster heads (CHs), can be physically compromised by a malicious adversary. By using compromised CHs, the adversary can intentionally attach false message authentication codes into legitimate sensing reports in order to interrupt reporting of the real events. The existing solutions are vulnerable to such a type of security attacks, called manipulated compilation attacks (MCAs), since they assume that CHs are uncompromised. Thus, the reports manipulated by compromised CHs will be discarded by forwarding nodes or rejected at base stations, so that real events on the fields cannot be properly reported to the users. In this paper, the author proposes a method for the detection of MCAs in cluster-based WSNs. In the proposed method, every sensing report is collaboratively generated and verified by cluster nodes based on very loose synchronization. Once a cluster node has detected an MCA for a real event, it can reforward a legitimate report immediately. Therefore, the event can be properly reported to the users. The performance of the proposed method is shown with analytical and experimental results at the end of the paper.