Spectrum sharing in cognitive radio systems: service-oriented capacity and power allocation

In this study, the authors investigate different capacity notions in cognitive radio systems where the transmission parameters of the cognitive users are adaptively changed based on the availability of (i) channel state information (CSI) pertaining to the cognitive (secondary) user link, and (ii) soft-sensing information (SSI) about the activity of the licensed-band (primary) user as obtained from the sensing detector at the secondary user's transmitter. Using statistics of available SSI and CSI at the secondary transmitter, the cognitive user adopts a transmission policy that maximises the achievable capacity under appropriate system constraints. Assuming above considerations in a cognitive radio system operating under average received-interference and peak transmit-power constraints, the authors study three different capacity notions of spectrum-sharing fading channels - namely, ergodic, delay-limited and service-rate with/without outage - and obtain their corresponding optimal power allocation policies. In particular, in this study, the authors obtain the optimal power allocation policies to achieve the channel capacity of the secondary link for the appropriate transmission policy which satisfies the spectrum-sharing system requirements in each of the above-mentioned scenarios, and further provide expressions for the capacity metrics. They also provide illustrative numerical simulations sustaining our theoretical results.