A Double Key-Sharing Based False Data Filtering Scheme in Wireless Sensor Networks

In wireless sensor networks, the attackers can inject false data reports containing bogus readings or non-existent events from compromising nodes, which may result in false alarms, interfere decision making, and drain out the limited energy of the network. Existing approaches to filter false reports, notably statistical en-route filtering, share keys between the nodes in low probabilities, and rely on the forwarding nodes to verify the correctness of the MACs (Message Authentication Codes) carried in each report. As a result, false reports have to travel multiple hops before being detected and filtered, and thus lead to the waste of energy of the network. A Double key-Sharing based false data Filtering scheme (DSF) in this paper is proposed to cope with this problem, in which nodes are grouped into clusters and each cluster head establishes relationship with the nodes closer through pair wise keys to form a blocked region. Furthermore, each pair of nodes shares symmetric keys randomly. When an event occurs, a legitimate report must carry two types of MACs. In filtering phase, each forwarding node validates not only the correctness of these two types of MACs carried in the report, but also drops part of the tail of the reports just outside the blocked region. Extensive analyses and simulations show that DSF outperforms existing schemes in terms of filtering efficiency and energy consumption.