Adaptive codebook precoder employing Kalman filter for time-varying MIMO channels

In this paper, an adaptive codebook precoding scheme is proposed for time-varying multiple-input multiple-output (MIMO) channels with feedback delay. The proposed scheme updates its codebook on the basis of the spherical cap codebook (SCC) approach. In contrast with the conventional radius calculation for SCC, the proposed scheme does not need additional feedback of the cap radius or offline radius calculation, but only adapts the radius according to a previously selected codeword. In addition, the Kalman filter using the differential model as its process equation is employed to predict MIMO channels and to select a suitable codeword from the codebook. Since the Kalman filter does not require information on the maximum Doppler frequency, the proposed scheme, which is a combination of the self-adaptive SCC and the Kalman filter, can work well even when the maximum Doppler frequency is not exactly known. Computer simulations under 4-by-4 MIMO channels with 2 data streams and a 6-bit codebook demonstrate that the proposed scheme can improve the system capacity in the same way as the conventional scheme does, which needs exact information on the maximum Doppler frequency in contrast to the proposed scheme. When the different channel model is used, it is shown that the proposed self-adaptive SCC can still provide considerable capacity gain, while the conventional scheme can not do so.