Skip to Main Content
Security is one of the most important properties of wireless sensor networks, which are threatened by varied attacks, to approach practical deployment. Intrusion detection technology is a positive method to defense against security threats in wireless sensor networks. In this paper, characteristics and attack model of wireless sensor networks are deeply studied, and an intrusion detection model is built to find out vicious nodes, namely, an improved BP algorithm is introduced, normal and misuse-based detection methods are set on cluster head node, anti-complement data are collected and analyzed by the rule base in real time. Simulation results show that the model is effective and accurate compared to traditional methods, and the performance of intrusion detection is improved.
Date of Conference: 27-29 June 2011