Performance bounds with precoding matrices compliant with standardized 5G-NR for MIMO transmission

Abstract

Advanced multiple-input multiple-output (MIMO) beamforming techniques are crucial in 5G New Radio (NR) to achieve the expected data rate values. Therefore, the 3rd Generation Partnership Project (3GPP) has proposed a codebook-based MIMO precoding strategy to provide high diversity, array gain, and spatial multiplexing. The main goal is to obtain a tradeoff between performance, signal overhead, and complexity. The precoding matrix is selected from a set of predefined codebooks based on the knowledge that the 5G-NR base station (gNB) acquires about the channel status. In this work, a detailed study of the precoding matrix design is provided following the guidelines reported in the technical specifications 38-211 and 38-214 of the 3GPP. An analysis of the performance in terms of spectral efficiency (SE) achieved by the 5G-NR precoding matrices is illustrated for a single-user MIMO scenario. These results are contrasted against the optimal singular value decomposition (SVD) solution in order to explore the gap between the standardized precoding proposal and the optimal one. Several values of signal-to-noise ratio (SNR) and different antenna array configurations are considered. Moreover, the multiplexing gain for a different number of parallel data streams is evaluated. Numerical results show the SE bounds that can be obtained with the 5G-NR precoding matrices. These insights are of key importance for practical implementation of precoding strategies in 5G-NR systems and beyond.