HMM based spectrum access options for cognitive radio networks

Cognitive radio networks have managed the radio spectrum more efficiently by using the idle spectrum in primary users´ licensed frequency bands. Recent researches tell us the spectrum utilization can be improved significantly if it is possible for cognitive users to access the free spectrum bands. In this paper, we formulate the spectrum access selection optimization problem as a hidden Markov model (HMM), which helps the cognitive users evaluate the channel quality. We will implement HMM predict next channel state based on past ones. Then we adopt the restless bandit method to obtain the optimal solution that makes better spectral efficiency and reduces the energy consumption.