Low-Complexity Precoding by Exploiting Spatial Sparsity in Massive MIMO Systems

Zhenkun Qiu,Shengli Zhou,Ming Zhao,Wuyang Zhou
DOI: https://doi.org/10.1109/twc.2021.3132789
IF: 10.4
2022-01-01
IEEE Transactions on Wireless Communications
Abstract:A massive multi-input-multi-output (MIMO) system can bring substantial improvement in spectral and energy efficiency for wireless communication systems. The high array gain and fine spatial resolution allow the utilization of relatively simple processing at the base station, such as the near-optimal regularized zero-forcing (RZF) precoding. Nevertheless, such precoding schemes require computing the inverse and multiplication of matrices, which leads to a large burden for baseband processing when the system dimension grows large. To take full advantage of the spatial sparsity of massive MIMO channels in the finite scattering environment, we propose a virtual channel precoding strategy that directly utilizes the sparse virtual channel representation (VCR) in precoding. Furthermore, we develop a virtual channel RZF precoding algorithm based on the pre-conditioned conjugate gradient method, namely PCG-VC-RZF, which can directly use the sparse virtual channel matrix while avoiding the complex computation of the Gram matrix. The computational complexity and transmission delay are shown to be reduced tremendously by the proposed algorithm when operating in a large-dimension system. We further provide a beam selection strategy for improving the performance of the proposed approach under a complexity constraint. The simulation results verify the efficiency of the proposed precoding scheme.
What problem does this paper attempt to address?