Optimal Cognitive Radio Spectrum Access With Joint Spectrum Sensing and Power Allocation

Abstract

Spectrum access policy in cognitive radio (CR) systems is classically based on a binary-valued function, where the secondary users (SUs) can access the spectrum of a primary user (PU) with the probabilities `0' or `1', according the measured energy. In this letter, following the concept of probabilistic spectrum access, which assigns a probability for CR signal transmission over the spectrum of a PU, we show that the optimal spectrum access approach belongs to the family of real-valued continuous functions where the probability of accessing or not accessing the spectrum by SUs is a real number between zero and one. Moreover, we show that the spectrum access probability is always greater than or equal to the tolerable miss detection probability. For comparison, we consider the well-known interweave (i.e., no transmission when PU is present) and underlay (i.e., always transmission even when PU is present) approaches where the spectrum access probability is either equal to `0' or `1'. Numerical results confirm that the proposed spectrum access outperforms the interweave and underlay schemes in terms of the achievable CR data rate.