Abstract
In this paper, we consider a heterogeneous cognitive radio (CR) system where a large number of secondary users (SUs), having different path losses, can share the spectrum with a primary user (PU). We assume that the secondary base station (SBS) has an imperfect channel state information (CSI) of the interference channel between the SBS and the PU and does not have the full CSI from SUs. We consider a two steps scheduling algorithm based on opportunistic beamforming (OB). In step one, the SBS generates orthogonal beams and transmits the generated beams to all SUs. Each cognitive user feeds back to the SBS its SINR (signal to interference and noise ratio) maximum and the index of the corresponding beam. The SBS selects the cognitive users having the maximum SINRs for transmission. In this paper, we study the impact of SINR quantization on the throughput of the secondary system based on the pdf-matched quantization for which the sum-rate distortion is minimized. We also investigate the fairness of our scheduling scheme. Simulation results show that our scheme achieves maximum throughput and fairness simultaneously