Multi-channel myopic sensing for opportunistic spectrum access using channel correlation

We consider the issue of opportunistic spectrum access (OSA) in cognitive radio network. By modeling the channel occupancy states induced by the primary users as a discrete-time Markov process, the channel selection scheme of the secondary users at the MAC layer can be formulated as a partially observable Markov decision process (POMDP). Under the POMDP framework, myopic sensing (which is a kind of channel selection scheme) is adopted with the assumption that occupancy of each channel evolves independently. In this paper, differing from former research assuming independent channel occupancy, we investigate into the case when correlation exists across different channels. By exploiting the channel correlation information, we propose two multi-channel myopic sensing policies: one with full-dimension belief vector, and the other with reduced-dimension belief vector. The sensing policies are given out under different PHY layer conditions: with or without the presence of spectrum sensing error. Simulation results illustrate that the proposed sensing policies can improve the throughput of the secondary users as time increases.