Deep Learning Detector for Large-Scale MIMO Systems with Low-Resolution ADCs

A large-scale multiple-input multiple-out (LS-MIMO) transmission scheme with low-resolution analog-to-digital converters (ADCs) has become one of the promising techniques for 5G and future wireless networks. In this paper, we investigate the power of a deep-learning network in detecting LS-MIMO signals when the resolution of the ADCs is limited to just a few bits. We found that the performance of the deep-learning detector is sensitive to the resolution of the input signals. And thus, it desires to train a specific deep-learning detector for each level of the resolution. Furthermore, the deep-learning detector can deliver equal or better performance than the belief propagation detector. At the high level of signal-to-noise ratio, the deeper the network is, the better performance of the detector is improved. This makes the deep-learning detector a promising technique to detect large-scale MIMO signals to achieve good performance while keeping the complexity manageable.