Abstract : Increased device connectivity and information sharingin wireless IoT networks increases the risk of cyber attack by maliciousnodes. In this paper, we present an effective and practicalscheme for detecting data integrity and selective forwarding attackslaunched by malicious relays in wireless IoT networks. Theproposed scheme exploits the broadcast nature of wireless transmissionand provides a sentinel based approach to intrusion detection.Our detection scheme assumes a general noise model for thenetwork where different wireless links may have different packeterror probability (PEP). Further, our detection scheme is effectiveeven in scenarios where different wireless links in the networkemploy distinct modulation and coding schemes at the physicallayer. This detection scheme has application in practical wirelessIoT networks, such as those based on the recently introduced IEEE802.11ah standard.
Index terms : Malicious relay detection, packet error probability, sentinel, wireless relay network.