Impact of Data Fusion on Intrusion Detection in Large-scale Wireless Sensor Networks

Intrusion detection is an important class of wireless sensor network applications with critical performance requirements. Due to the high deployment cost, it is crucial to understand and predict the intrusion detection capability of a surveillance sensor network. However, most existing analytical studies are based on simplistic sensing models (e.g., the disc model) that do not capture the stochastic nature of detection. In practice, data fusion has been adopted in a number of sensor systems to deal with sensing uncertainty and enable the collaboration among sensors. However, the performance analysis of wireless sensor networks designed based on data fusion has received little attention. In this paper, we investigate the fundamental detection performance of large-scale sensor networks under stochastic sensing models that take into account several realistic sensing characteristics such as signal decay and false alarms. Our results show that 1) data fusion is effective in achieving stringent performance requirements such as short detection delay and low false alarm rates, especially in the scenarios with low signal-to-noise ratios (SNRs); 2) the disc sensing model is only suitable when the SNR is sufficiently high. Our results help understand the impact of data fusion and provide important guidelines for the design of wireless sensor networks for intrusion detection.