A Unified Framework for Precoding and Pilot Design for FDD Symbol-Level Precoding

Large-scale antenna array techniques are key enablers for modern wireless communication systems. Channel state information (CSI) is indispensable large-scale multi-antenna systems, but challenging to obtain. To tackle this issue, in paper we propose a unified precoding and pilot design frame-work, that allows minimal precoding-sensitive modified CSI (mCSI) be collected. This results significant reduction the overheads complexity compared classical physical (pCSI) estimation. Based on framework, further an intelligent (IP) approach senses selects mCSI The IP utilizes compressive sensing formulation attach selection of optimization. We apply above multi-user frequency division duplexing (FDD) downlink as example. Our study shows advantages three-fold. First, contrast pCSI, only captured, which reduces training feedback overheads. Second, precoders optimized directly based mCSI, avoids recovering pCSI high-dimension. Third, since reduced dimension utilized, scale problem optimize precoder also thus it much easier solve.