Optimization of feedback for adaptive MIMO transmissions over time varying channels

This paper studies the capacity issue of adaptive multiple-input-multiple- output (MIMO) transmissions with channel state information (CSI) feedback over a time-varying channel which is modeled as time-correlated blocks. The CSI is estimated at the receiver and fed back periodically to the transmitter to enable adaptive transmissions. With the objective to maximize the ergodic capacity, a CSI-feedback strategy is proposed based on the optimization of following two parameters. (1) The bit number for the quantization of each CSI and (2) the time interval to refresh the CSI. The first parameter determines the accuracy of the quantization and the latter one deal with the time variation of the channel. The optimization is conducted under a given feedback-capacity constraint i.e., the maximum number of feedback bits per block is fixed in average. A theoretical result is derived and numerical results are presented to show the effectiveness of the proposed strategy.