Combined antenna selection and beamforming in cross-layer design for cognitive networks

We investigate the performance of an antenna selection algorithm applied with precoding to improve the throughput performance of cognitive multiple-input multiple-output (MIMO) radios. We assume cognitive MIMO radios coexist in the same frequency band with the primary users. In such event, for concurrent spectrum access, cognitive nodes use beamforming techniques to cancel the mutual information between cognitive and primary users. In that, cognitive nodes exploit the degrees of freedom offered by MIMO systems to beamform the transmitted signal in an attempt to cancel interference at primary users. In the cross-layer design, cognitive nodes maximize an objective function for the link layer throughput where precoding is applied at the physical layer on the transmitted spatial multiplexed signals. We present a closed form solution for the overall throughput in-terms of the physical layer parameters. Numerical results are presented to show the efficacy of the proposed scheme for different network settings.