Adaptive modulation for MIMO systems and throughput evaluation with realistic channel model

Abstract

MIMO systems exploit the multipath propagation in rich scattering environment using multiple transmit and receive antennas to increase the capacity of a link. The matrix channel plays a pivotal role in the throughput of a MIMO link since the modulation, data rate, power allocation and antenna weights are dependent on the channel gain. This paper evaluates adaptively modulated open loop MIMO system with limited feedback and adaptively modulated closed loop MIMO systems. For open loop MIMO a modified V-BLAST system with limited feedback is considered. This is compared with a rate maximization algorithm for closed loop MIMO using iid flat-fading Rayleigh channel, spatially correlated channel and realistic indoor channel. It is shown that when realistic models are used the difference in performance between closed loop MIMO and limited feedback V-BLAST systems is greater. At a noise-normalized power of 25 dB the percentage improvement in the throughput of CL-MIMO over OL-MIMO is 15.1% for Rayleigh fading channel, 48.1% for the spatially correlated channel and 104% for the realistic channel model. It is also shown that the throughput reduces in presence of channel correlation and that this affects the performance of the limited feedback V-BLAST system to a greater extent in comparison with CL-MIMO system. The average SNR penalty for OL-MIMO to achieve same throughput as CL-MlMO is 4dB for the Rayleigh fading model, 7dB for the spatially correlated model and 10 dB for the realistic channel model.