Low Complexity Hybrid Precoding for Millimeter Wave MIMO Systems

Abstract

Massive multiple input multiple output (MIMO) technology and millimeter wave (mmWave) system, as the key technologies of the new generation of mobile communications, can effectively increase channel capacity and relieve spectrum resources. Because the mmWave has a short wavelength, the transceiver can be composed of a large antenna array to reduce severe signals attenuation. Furthermore, the use of hybrid precoding technology can improve system performance and reduce system hardware complexity. The classic hybrid precoding algorithm that based on simultaneous orthogonal matching pursuit (SOMP) requires matrix inversion, which leads to high complexity, and its performance depends on the accuracy of channel estimation. In this paper, by modeling the mmWave MIMO system, we compare three improved algorithms, which are orthogonality based matching pursuit algorithm (OBMP), matrix-inversion-bypass simultaneous orthogonal matching pursuit algorithm (MIB-SOMP) and residual matrix-singular value decomposition algorithm (RM-SVD). We analyze the performance of the algorithms, such as complexity, spectrum efficiency, bit error rate, as well as the advantages and disadvantages of the algorithms.