Precoder selection and rank adaptation in MIMO-OFDM

The knowledge of the channel state information (CSI) at the transmitter can significantly improve the performance of a MIMO system. In a limited-feedback MIMO system, the feedback overhead from the receiver to the transmitter is significantly reduced by using a codebook. The codebook is made up of finite set of precoding matrix that is predetermined by quantizing the CSI. The receiver selects the best precoding matrix and sends back only the index of the selected precoding matrix. However, in MIMO-OFDM if the precoding is performed per subcarrier, the feedback overhead still remains large. Therefore, to reduce the amount of feedback one precoder per a group or band of subcarriers is used. In this paper we develop reduced-complexity algorithms that select the best precoding matrix per band of subcarriers. We also derive rank adaptation algorithm for further performance improvement.