Mitigating the Effect of Malicious Users in Cognitive Networks

Abstract

Open medium access is an inherent feature of wireless networks that makes them vulnerable to security threats and unauthorized access by foreign entities and malicious users. In cognitive radio networks, effect of these malicious users on network performance becomes more threatening, where they impersonate primary users by emitting similar signals, causing secondary users to vacate the occupied channel needlessly. As a result, network resources are unfairly monopolized by malicious users denying other secondary users their fair share. In this paper, we introduce a cross-layered approach to provide secondary users the ability to differentiate between a primary user and a malicious user, using Hidden Markov Model at MAC layer. Hence, our proposed framework allows the transport layer protocol to respond appropriately in a way that the effect of the presence of malicious user on the network is mitigated. The effectiveness of our proposed approach is shown by calculating the throughput of the network and number of channel switches with respect to varying number of secondary and malicious nodes and by comparing it to an earlier proposed TCP protocol.