Reinforcement learning method for energy efficient cooperative multiband spectrum sensing

Cognitive radios (CR) and dynamic spectrum access (DSA) attempt to exploit the underutilized radio spectrum by allowing secondary users to access the licensed frequencies in an opportunistic manner. In order to avoid collisions with the primary user the secondary users need to sense the spectrum, and to mitigate the effects of channel fading on sensing cooperative schemes have been proposed in the literature. In this paper a multiband spectrum sensing policy for coordinating cooperative sensing is proposed. The proposed policy employs the ϵ-greedy reinforcement learning method to prioritize the sensing of different subbands and to assign those secondary users to sense them that are able to provide a desired level of miss detection probability. In order to improve the energy efficiency, the number of assigned sensors per subband is minimized.