State Estimation and Anomaly Detection in Wireless Sensor Networks

Abstract

Reliable co-operative wireless sensor networks are now used in a variety of disciplines including but not limited to geoscience, medical science and security management. Sensor nodes are low-powered microelectronic devices with limited communication and sensing range. In a cluster-based network, sensor nodes are grouped into a number of clusters based on their physical locations. Sometimes, an anomalous node is cleverly placed into the network by intruders for reducing power and efficiency of the network. These nodes collect and send the confidential information to outsiders. Additionally, an anomalous node can destroy the network by gradually damaging inter-node dependence within a cluster. We review some existing and recent statistical models for locating such anomalous node and recovering efficiency of the network. Then, we propose a novel dynamic linear mixed model for the simultaneous state estimation and anomaly detection. Our proposed model can efficiently locate an anomalous node in a relatively short time and the power of the model is evaluated by numerical studies. The proposed approach will be very useful in medical science, military surveillance and environmental studies.