Capacity of Cognitive Radio with Partial Channel Distribution Information in Rayleigh Fading Environments

Abstract

AbstractThis paper investigates the capacity of the secondary user (SU) in a cognitive radio (CR) network in Rayleigh fading environments. Different from existing works where perfect channel state information (CSI) or channel distribution information (CDI) of the interference link from the SU to the primary user (PU) is assumed to be available, this paper assumes that only partial CDI is available. Specifically, we assume the distribution parameter is unknown and estimated from a set of channel gain samples. With such partial CDI, closed-form expressions for the ergodic and outage capacities of the SU are obtained under the transmit power and the interference outage constraints. It is shown that the capacity with partial CDI is not degraded compared to that with perfect CDI if the interference outage constraint is loose. It is also shown that the capacity can be significantly improved by increasing the number of channel gain samples.