CSI-aided MAC with Multiuser Diversity for Cognitive Radio Networks

Abstract

Cognitive Radio (CR) aims to increase the spectrum utilization by allowing secondary users (SU) to access unused licensed spectrum bands. To maximize the throughput given limited sensing capability, SUs need to strike a balance between sensing the channels that are not heavily used by primary users (PU) and avoiding collisions with other SUs. To randomize sensing decisions without resorting to multiuser sensing policies, it is proposed to exploit the spatially-variant fading channel conditions on different links by adapting the reward to the channel state information (CSI). Moreover, the proposed channel-adaptive policy favors links with high achievable transmission rate and thus further improves the network throughput.