End-to-end approach for MU massive MIMO uplink transmission under 1-bit ADC

Abstract

Multi-User (MU) Massive Multiple Input Multiple Output (MIMO) is considered a potential and disruptive technology for fifth-generation (5G) and sixth-generation (6G) wireless systems. The presence of antennas networks at the base station level to communicate with single antenna users leads to in excessively high system costs and power consumption. The deployment of the analog-digital converters (ADC) 1 bit in the BS can solve these problems. In this work, we study the compensation of the impact of 1 bit ADC in MU massive MIMO systems on the uplink transmission accompanied with Minimum Mean Square Error (MMSE) detector. First, we propose a new end-to-end (E2E) learning approach based on Deep neural network (DNN). The objective is to design the communication system as a set of layers for the transmitter and the receiver in unknown channels. Then we present a mixed ADC architecture, where some antennas use 1-bit ADCs while the rest does not exploit ADCs using the end-to-end approach. The simulation results demonstrate the high potential of the proposed E2E approach in terms of uplink quality improvement. It is also demonstrated that the performance of massive MIMO systems with a mixed ADC architecture approaches those of ideal systems.