The hardware cost and power consumption of a massive multiple-input multiple-output (MIMO) system can be remarkably reduced by using a very low-resolution analog-to-digital converter (ADC) unit in each antenna. However, such a pure low-resolution ADC architecture complicates parameter estimation problems. These issues can be resolved and the potential of a pure low-resolution ADC architecture can be achieved by applying a mixed ADC architecture, whose antennas are equipped with low-precision ADCs, while few antennas are composed of high-precision ADCs. In this paper, a unified framework is presented to develop a family of detectors on a massive MIMO uplink system through probabilistic Bayesian inference. Our basic setup comprises an optimal detector, which is developed to provide a minimum mean-squared-error estimate on data symbols. Considering that highly nonlinear steps are involved in quantization, we also investigate the potential for complexity reduction on an optimal detector by postulating a common pseudo-quantization noise model. We provide asymptotic performance expressions, including mean squared error and bit error rate for optimal and suboptimal MIMO detectors. These expressions can be evaluated rapidly and efficiently. Thus, they can be used for system design optimization.

View PDF No PDF? Try Arxiv!