A Novel Limited Feedback Approach Based on Prediction for MIMO Beamforming System

This paper proposes a novel limited feedback approach based on prediction for MIMO (multi-input multi-output) beam forming systems over temporally-correlated channels. In the proposed approach, the receiver reports a codeword representing CSI (channel state information) back to the transmitter at a regular interval of several data frames, thereby the feedback load is remarkably reduced. As temporal correlation across adjacent frames, the AR (auto-regressive) prediction model is exploited to compensate for performance impairment due to feedback delay. Selection of the representational beam forming codeword is optimized based on a cost function that involves CSI of all frames. The simulation results show that, the proposed approach can reduce efficiently the feedback rate meanwhile maintaining good BER (bit error ratio) performance.