Channel Aware Detection of Selective Forwarding Attacks in Wireless Sensor Networks

Abstract

Wireless Sensor Networks (WSN) are widely deployed in many applications for their effective event monitoring and data gathering capabilities, so they may be susceptible to Selective Forwarding Attacks in which some of the forwarding packets may be dropped maliciously to reduce the performance of the network. Due to the changing conditions of wireless channel, the packet loss rate may be high and vary with time. So, it poses a problem in distinguishing malicious drops from traditional packet loss. This paper proposes an efficient method to detect Selective Forwarding Attacks in Wireless Sensor Networks by penalizing the nodes that are identified to be compromised. To improve detection accuracy, an optimal threshold that is adaptive to the time-varied channel condition is derived. The attack probabilities of malicious compromised nodes is also estimated. The proposed system can identify compromised sensor nodes. It also improves the packet delivery ratio of the network.