Downlink Channel Estimation in Massive MIMO FDD Systems Using Block-ADMM

The fifth-generation (5G) wireless communication systems envision the deployment of massive multiple-input multiple-output (MIMO) technology for enhanced network coverage and capacity. With frequency division duplex (FDD) being the prominent mode of operation, addressing the challenges for massive MIMO systems in FDD is extremely important. One such challenge is the downlink channel estimation, which should be accomplished with minimal pilot overhead. To address this challenge, we propose a compressive sensing (CS) algorithm based on alternating direction method of multipliers (ADMM). We demonstrate how our proposed algorithm fully exploits the block sparsity inherent in the spatial correlations of MIMO channels to provide improved performance than the conventional CS-based algorithms used for massive MIMO FDD systems.