Interference-Aware Power Allocation in Cognitive Radio Networks with Imperfect Spectrum Sensing

This paper investigates the problem of downlink power allocation in a Cognitive Radio (CR) network that operates by opportunistically utilizing vacant licensed Primary User (PU)spectrum, where it is assumed the PU employs an OFDMA scheme. We first analyze the interference to the PU caused by the CR network due to imperfect spectrum sensing and out-of-band (OOB) emissions from CR subcarrier side-lobes. The power allocation algorithm aims to maximize the CR network capacity under interference constraints, while considering the effect of spectrum sensing errors such as false alarm and misdetection probabilities. The power allocation solution for a single-user CR network is the waterfilling criteria with a variable water level in different subcarriers. For a multiple-user CR network, we employ a heuristic algorithm to solve the power allocation problem, with a significantly reduced computational complexity as compared to the optimal solution.