Efficient Iterative Linear Precoding Scheme for Downlink Massive MIMO Systems

Massive multiple-input multiple-output (MIMO) is the crucial technology to increase the 5G wireless communication system’s reliability and throughput. Massive MIMO uses a combination of a precoder and massive antennas at the base station (BS). A simple beamforming strategy such as zero-forcing (ZF) can be exploited in massive MIMO. ZF is a linear precoding technique usually adopted in low complexity massive MIMO systems. However, ZF precoding techniques involve matrix inversion, whose size increased with the user equipment. It increases the system’s overall computational complexity. In this paper, a modified weighted two-stage (WTS) algorithm is proposed to minimize that effect. The existing WTS algorithm used two symmetric half iterations and combined the iterations for faster convergence. It demands computation in the forward and the reverse order in each iteration. However, the proposed modification considers the present and past iterations, eliminating the reverse iterations and minimizing the complexity. The proposed change reduces the computational complexity by 17%. Simulation results show that modified WTS achieves the near-optimal capacity and similar bit error rate (BER) performance as ZF precoding.