Distinguishing Data Transience from False Injection in Sensor Networks

Wireless sensor networks are increasingly being employed for potentially hazardous and critical applications such as monitoring the gas concentration levels in a battle held. In such sensitive applications, it is vital to monitor closely the transient phenomenon and take requisite preventive and corrective actions, if necessary. In such scenarios, due to the presence of adversaries who intend to disrupt the functioning of the system, it becomes imperative to shield our system from false data injection attacks. We propose a novel secure statistical scheme, called SSTF to effectively monitor the transient phenomenon while being immune to false data injection attacks. For achieving our goals, we require the sensors to do a lightweight computation and report a simple statistical digest in addition to the current sensed reading. SSTF is a two-tier system consisting of a statistical inter-sensor testing framework, which is the kernel of our scheme, employed atop an enhanced version of a well-known existing security scheme. We present detailed theoretical analysis and in- depth simulations to show the effectiveness of SSTF.