Multiple Dimensional Scaling Hybrid Precoding in Millimeter Wave MIMO System

Abstract

AbstractIn most wireless systems, the channel transmission loss of millimeter wave signals is often greater than that of traditional signals. Hybrid precoding can leverage large antenna arrays to compensate this loss and further improve the spectral efficiency of millimeter wave. This paper explores the sparsity of mmWave channel—only a few main paths are useful for the precoding procedure. This sparsity makes it possible to use a limited feedback in channel estimation. We propose a novel hybrid precoding scheme in this paper, which utilizes the main vector of the channel. We call this scheme Multiple Dimensional Scaling (MDS) hybrid precoding. On the one hand, comparing with many traditional hybrid precoding schemes with full channel information feedback, our scheme can decrease feedback overheads significantly. On the other hand, comparing with some hybrid precoding with limited channel feedback information, our scheme can improve spectral efficiency. Moreover, the simulation results show that the system spectral efficiency increases with the number of data streams significantly.KeywordsMillimeter wave communicationsHybrid precodingLimited feedbackMultiple Dimensional Scaling