A Compressed Analog Feedback Strategy for Spatially Correlated Massive MIMO Systems

In multiuser MIMO systems, the amount of the required feedback increases with the number of transmit antennas to improve the sum rate performance. This feedback overhead is critical when it comes to massive multiple antenna wireless systems. Therefore, we propose a compressed analog feedback scheme. The proposed compressed analog feedback strategy exploits natural correlation characteristics among large-scale multiple antennas in order to reduce feedback overhead. By using the sparse approximation and low dimensional random projection at each mobile station (MS), and convex relaxation based feedback decompression at a base station (BS), the proposed scheme outperforms the uncompressed analog feedback which wastes long time duration and energy, in the consideration of the feedback delay, requiring lower feedback overhead from each MS.