Distributed T-Distribution-Based Intrusion Detection in Wireless Sensor Networks

Detecting malicious attackers is a critical problem for many sensor network applications. In this paper, a distributed t-distribution-based intrusion detection scheme was proposed. Considering the spatial correlation in the neighborhood activities, our intrusion detection algorithm established a robust model for multiple attributes of sensor nodes using t-distribution. The robust model with an approximate parameter algorithm was exploited to detect malicious attackers precisely. Experimental results show that our algorithm can achieve high detection accuracy and low false alarm rate even when a few sensor nodes are misbehaving, and perform quickly with a lower computational cost.