Adaptive codebook for limited feedback MIMO system

Abstract

In multi-user MIMO system, full multiplexing gain relies on accurate channel knowledge provided at the transmitter, but usually it cannot be directly acquired by the transmitter. One solution is the limited feedback scheme where each user quantizes its channel with a fixed codebook and sends it back to the transmitter. There are two types of codebooks frequently discussed in IEEE 802.16e: 16e codebook and DFT codebook. Due to different design criterions, 16e codebook achieves a better performance in uncorrelated channel and DFT codebook is more suitable for correlated channel. Since the real environment is a combination of uncorrelated channel and correlated channel, we propose an adaptive codebook which adapt to different channel distribution. Our adaptive codebook is constructed with both 16e and DFT codebooks, where DFT codebook is employed to reconstruct the channel spatial correlation matrix and shift 16e codebook to fit the current channel. This adaptive codebook obviously outperforms both 16e and DFT codebooks in any channel distribution with only a negligible increase in overhead. Moreover, our adaptive codebook avoids complicated mathematic operation, which is very easy to be implemented in practice.