Feedback bit allocation for large distributed antenna systems in indoor WLAN

Abstract

Massive MIMO is a key technique for next generation wireless networks due to its potential for significant capacity improvement. To achieve good performance, accurate channel state information (CSI) is needed, especially at the transmitter, to calculate the beamforming matrices. This problem exists for traditional MIMO systems and becomes more challenging for Massive MIMO as the number of antennas is much larger. In this paper, we focus on Massive MIMO with distributed antenna system (DAS) architecture, which we refer as large DAS. In a large DAS, a large number of remote antenna units (RAUs), each equipped with a small antenna array, are distributed over the service area, cooperatively working for beamforming. Due to the large separation of the RAUs, the large scale fading factors of the RAUs for a given mobile station (STA) is different. This effect can be utilized for optimizing the feedback bit allocation considering that more bits should be allocated to dominant signal paths. So we propose two bit allocation algorithms, one that adaptively allocates different bits to the RAUs according to the large scale fading, termed adaptive allocation; and the other that allocates equal bits to a set of RAUs, termed equal allocation. They are applied for centralized and de-centralized zero-forcing beamforming (C- and D-ZFBF) respectively. The results show that with a proper bit allocation, the data rates can approximate that of perfect CSI with limited feedback rate. For C-ZFBF, adaptive allocation performs better than equal allocation with medium feedback rates, and similarly for low and high feedback rates. For D-ZFBF, adaptive allocation and equal allocation offer similar results.