Suboptimal Approach to Distributed Detection in Cognitive Radio Networks

In this paper a dynamic spectrum access (DSA) concept is explored for mitigating the paucity of spectral bandwidth in cognitive radio (CR) for opportunistic, dynamic access of the spectrum without any interference. Dynamic spectrum access schemes are proposed for a distributed cognitive radio network consisting of one secondary user (SU) and many primary users (PUs). The SU has to make decisions for accessing PU channels within discrete time slots. The design of sensing and access strategies that govern channel choice in each slot for near-optimal throughput performance of the SU may be formulated as a partially observable Markov decision process (POMDP). Furthermore, it is considered that the SU incurs a cost whenever it switches to a different channel. The switching cost is expressed in terms of delay, packet loss and packet overhead. In this work, the SU access policy based on a myopic approach is proposed and evaluated.