In this article, deep learning is applied to estimate the uplink channels for mixed analog-to-digital converters (ADCs) massive multiple-input multiple-output (MIMO) systems, where a portion of antennas are equipped with high-resolution ADCs while others employ low-resolution ones at the base station. A direct-input deep neural network (DI-DNN) is first proposed to estimate channels by using the received signals of all antennas. To eliminate the adverse impact of the coarsely quantized signals, a selective-input prediction DNN (SIP-DNN) is developed, where only the signals received by the high-resolution ADC antennas are exploited to predict the channels of other antennas as well as to estimate their own channels. Numerical results show the superiority of the proposed DNN based approaches over the existing methods, especially with mixed one-bit ADCs, and the effectiveness of the proposed approaches on different ADC resolution patterns.
@article{arxiv.1908.06245,
title = {Deep Learning based Channel Estimation for Massive MIMO with Mixed-Resolution ADCs},
author = {Shen Gao and Peihao Dong and Zhiwen Pan and Geoffrey Ye Li},
journal= {arXiv preprint arXiv:1908.06245},
year = {2019}
}
Comments
This paper has been accepted by IEEE Communications Letters