Two new power allocation schemes for an OFDM cognitive radio with no knowledge on primary users’ interference

In this paper, we propose two new power loading schemes for an OFDM cognitive radio (CR). The novelty of this paper is that the CR has no knowledge about the interference power introduced by primary users (PUs). In the first scheme, we maximize the total received signal power on all CR subcarriers as the objective function with the two constraints: (i) keeping the interference to the PUs below a given threshold, and (ii) allocating less than peak transmit power as the total power levels on all CR subcarriers. The second proposed scheme is similar to the first one with one more constraint, which is more power loading to the CR subcarriers experiencing higher ratio of channel power gain to the noise power. This third constraint results in increasing CR throughput (similar to the water‐filling scheme). Performances of these schemes are analyzed by numerical and simulation results, which illustrate that the two proposed schemes achieve relatively close to the optimal scheme (in which the secondary transmitter has full knowledge about PUs’ interference power) and outperform the scheme based on estimation of this interference power with an estimation error higher than a given threshold. Copyright © 2012 John Wiley & Sons, Ltd.