Performance Improvement of Opportunistic Beamforming with Conservative Link Adaptation in the Presence of Channel Estimation Error

Opportunistic beamforming (OBF) offers a way to provide the multiuser diversity in slow fading channel by using randomly generated beam weights, leading to the substantial reduction of the required feedback from users. In spite of the advantage of the reduced feedback, the imperfect channel estimation might influence the quality of the estimated SNR and channel scheduler so bad that the selected AMC level would be higher than the achievable rate of the actual channel, which is called outage. In this paper, we propose a conservative link adaptation, where the estimated SNR is scaled down by a conservative factor which minimizes the variance of the difference between the actual channel SNR and the resultant SNR. Simulation results show that the introduction of conservative factor shows two-fold performance improvement in the presence of channel estimation error