Localization Anomaly Detection Algorithms For Wireless Sensor Networks

In wireless sensor networks (WSNs), the locations of sensors play a critical role in many applications, such as environment monitoring, geographical routing, target tracking and so on. However, the error locations may be generated because of the adversaries’attacks and bring severe influence on many applications based on the locations of sensors. Therefore, secure localization have become a research hotspot now. And localization anomaly detection as a kind of active defense technology against attacks is an important research direction of network security.We first propose a localization anomaly detection algorithm based on deployment model named BMLAD. Sensor nodes first acquire the actual number of neighbor nodes, then calculate the estimated number of neighbor nodes according to the deployment model knowledge and locations of sensors. The estimated and actual number of neighbor nodes form difference matrix (DM). Sensor nodes judge the locations by comparing the difference matrix and the threshold. If the value of DM is larger than a threshold value, we say the location is abnormal.We need to know the deployment model of sensor network in BMLAD, but not all of the deployment model of sensor networks can be obtained. So, we introduce clustering network and propose a localization anomaly detection algorithm based on clustering named BCLAD. After deploying all of the sensor nodes, sensor nodes begin to cluster according to LEACH protocol. Each cluster head node judged whether the sensor nodes are abnormal according to the location and neighbor list from sensor nodes.In this paper, theoretical research and simulation experiments are carried out for BMLAD and BCLAD. The simulation results show that BCLAD is applicable to the sensor network whose deployment model is unknown and BMLAD is applicable to the sensor network whose deployment model is known. In BCLAD and BMLAD, the smaller the abnormal node ratio, the higher the detection rate.