Statistical Channel State Information Aided Proportional Fair Scheduling Scheme for Highly Transmit Correlated Channels

In this paper, we consider the downlink heterogeneous scenario with highly transmit correlated channels, and propose two kinds of the statistical channel state information (SCSI) aided proportional fair scheduling (PFS) schemes: (1) signal-to-noise ratio based PFS (SNR PFS) scheme and (2) SNR PFS scheme with threshold, which determine the preferred beamforming vector and proportional fair index by exploiting the SCSI and only send back the instantaneous channel quality indicator (CQI) without the corresponding precoding matrix index at each timeslot. Although the SNR PFS scheme only requires little feedback overhead and low computational complexity, the performance of the SNR PFS scheme is nearly the same as the conventional scheme (i.e., codebook-based PFS scheme). However, with the increasing number of users, a large number of CQI feedbacks will increase the signaling burden. To overcome this obstacle, the proposed SNR PFS scheme with threshold which sets an appropriate feedback threshold to limit the CQI feedbacks with poor channel quality would hardly affect the performance of the scheme. We further derive the approximate rate expressions of the SNR PFS scheme by using probability statistics and the theory of order statistics, whose result is nearly approximated to that of the Monte Carlo simulation. Finally, our simulation results show that the performance of the SNR PFS scheme with βth =  ln(K/5) is close to that of the SNR PFS scheme, while it has less limited feedback capacity than the SNR PFS scheme.