Anti-jamming communication in cognitive radio networks with unknown channel statistics

Recently, many opportunistic spectrum sensing and access protocols have been proposed for cognitive radio networks (CRNs). For achieving optimized spectrum usage, existing solutions model the spectrum sensing and access problem as a partially observed Markov decision process (POMDP) and assume that the information states and/or the primary users´ (PUs) traffic statistics are known a priori to the secondary users (SUs). While theoretically sound, these existing approaches may not be effective in practice due to two main concerns. First, the assumptions they made are not practical, as before the communication starts, PUs´ traffic statistics may not be readily available to the SUs. Secondly and more seriously, existing approaches are extremely vulnerable to malicious jamming attacks. A cognitive attacker can always jam the channels to be accessed by leveraging the same statistic information and stochastic dynamic decision making process that the SUs would follow. To address the above concerns, we formulate the problem of anti-jamming multichannel access in CRNs and solve it as a non-stochastic multi-armed bandit (NS-MAB) problem, where the secondary sender and receiver adaptively choose their arms (i.e., sending and receiving channels) to operate. The proposed protocol enables them to hop to the same set of channels with high probability in the presence of jamming. We analytically show the convergence of the learning algorithms, i.e., the performance difference between the secondary sende and receiver´s optimal strategies is no more than O(20k/√ε √Tn ln n). Extensive simulations are conducted to validate the theoretical analysis and show that the proposed protocol is highly resilient to various jamming attacks.