Fast Iterative Hybrid Precoding and Combining With Momentum Gradient Descent and Newton’s Method for Millimeter Wave MIMO Systems

Abstract

Millimeter-wave (mmWave) massive multiple-input multiple-output (MIMO) systems have attracted much attention from both researchers and industry professionals, as they are seen as a suitable solution to the growing demand for cellular services in fifth-generation (5G) and sixth-generation (6G) wireless communication systems. In mmWave MIMO systems, iterative hybrid precoding/combining algorithms, which use a combination of analog and digital precoders, have gained significant interest because they perform comparably to fully digital precoding/combining while operating at reduced complexity due to a lower number of radio frequency (RF) components. However, the problem with these algorithms is that their convergence requires a substantial number of iterations. This paper solves this problem and introduces a fast convergence iterative hybrid precoding/combining algorithm using momentum and Newton’s method (FIHB-MN) for mmWave MIMO systems. Simulation results demonstrate the faster convergence of the algorithm’s objective function compared to other iterative methods in the literature. Moreover, FIHB-MN provides performance similar to unconstrained digital precoding and combining with only a few iterations. The simulation results also confirm that the spectral efficiency as well as the bit error rate (BER) performances of the proposed FIHB-MN algorithm outperforms other hybrid beamforming methods in the literature, all while maintaining low computational complexity.

Keywords

• mmWave MIMO systems
• hybrid design algorithms
• momentum gradient descent
• Newton’s method