A game theoretic approach for robust power control in cognitive radio networks

In cognitive radio networks, it is challenging for secondary users (SUs) to keep track of their interference at the receivers of primary users (PUs), due to the error in channel estimation and irregular information exchange between SUs and PUs. In this paper, we practically consider that SUs have only partial knowledge about the channel gains from SUs to PUs, based on which SUs estimate the worst-case channel gains and decide transmit power to robustly protect PUs. As it is rare that all SU-PU channels experience the worst-case conditions simultaneously, we proposed the worst-case selective robust model for SUs to estimate the aggregate interference power at PU receivers by predicting that only a part of SU-PU channels are in the worst-case conditions. We study SUs´ robust power control problem in a non-cooperative game, where each SU maximizes its own throughput subject to interference constraints at PU receivers. We propose an iterative algorithm for SUs to achieve unique Nash equilibrium in a distributed manner. Extensive numerical results show that our algorithm provides guaranteed protection for PUs provided with uncertain SU-PU channel information.