Multi-user Detection with Oversampled Large Antenna Arrays and Low-resolution ADCs

This paper delves into the uplink scenario of mmWave massive MIMO systems with dense ULAs and low-resolution ADCs. We focus on reducing power consumption and simplifying hardware while enhancing quantized system performance through spatial oversampling. We address the emergence of spatial thermal noise correlations and hardware imperfections, which profoundly affect signal recovery. To combat these challenges, we propose a non-linear inference method based on Vector Approximate Message Passing (VAMP) and Belief Propagation. Our algorithm aims to reconstruct transmitted signals from quantized measurements obtained by coupled antennas. We demonstrate that employing oversampling techniques significantly improves system performance, even when oversampled, highlighting spatial oversampling as an effective strategy for enhancing low-resolution ADC performance. Additionally, we analyze the impact of noise figure on recovery, underscoring its importance in system design.