Low Complexity Hybrid Precoding for Broadband mmWave Massive MIMO Systems

Abstract

Hybrid precoding becomes a candidate for Millimeter wave (mmWave) massive MIMO (Multiple-Input and Multiple-Output) systems because it can extremely reduce power consumption and high costs. Most prior work considered hybrid precoding for narrowband systems. However, wideband systems with frequency selectivity are likely to be operated in the future. In broadband systems, a common analogue precoder is designed for the overall frequency band whereas different digital precoders are employed in different subcarriers. In this paper, we propose the hybrid precoding schemes for broadband mmWave massive MIMO systems. First, the hybrid single-user (SU) algorithm is proposed for a single-user system. The common analogue precoding matrix is derived from the Equal Gain Transmission (EGT) method and the digital precoding matrices for different subcarriers are employed based on directly water-filling technique. Second, the hybrid multi-user (MU) algorithm is proposed for a multi-user system. Gram-Schmidt orthogonalization is added in the analogue domain and zero-forcing (ZF) is utilized for digital precoders in order to nullify inter-user interference. Simulation results show that our proposed hybrid schemes with low complexity can almost reach the performance of fully digital precoding algorithm and outperform other hybrid algorithms.

Keywords

• millimeter wave communication
• MIMO
• signal processing
• wireless communication
• wideband