MMSE algorithm based two stages hybrid precoding for millimeter wave massive MIMO systems

Abstract

Millimeter wave (mm-wave) communications has been sighted as solution becoming strong technology for the fifth generation networks. That is because, it brings orders of degree higher spectrum than current cellular bands and helps applications that require weak latency. Massive MIMO is used to minimize the severe constraints enjoined by the excessive propagation loss for millimeter wave frequencies. Where, normal implementation of antenna arrays is transmitted to multiple users simultaneously. Due to hardware constraints in mm-wave systems, classical lower frequency precoding techniques are difficult to be applied at mm-wave. A proposed modified hybrid precoding is suggested for downlink massive MIMO mm-wave systems in this paper. The hybrid precoding consists of mixed of analog and digital processing which is based on minimum mean square error (MMSE) algorithm. Where, quantized beamstearing codebooks are used to produce the analog and digital beamforming vectors with small training and feedback overhead. Extensive simulation programs are executed to compare the performance of the proposed precoding scheme with analog beamforming performance at different digital baseband precoders such as zero forcing precoder. Moreover, studies on the impact of modified hybrid beamforming parameters on the system performance are done. These parameters such as; the number of BS antennas, the number of mobile station antennas, number of effective channel quantization bits and number of RF beamforming quantization bits. Besides, the impact of rate threshold on the coverage probability is studied under different values of NBS antennas. It is shown that the performance of the modified hybrid precoding using MMSE is very close to the single user performance with only 0.5 b/s/Hz performance degradation at SNR = − 5 dB.