FFDNet Based Channel Estimation for Multiuser Massive MIMO System with One-Bit ADCs
Md. Habibur Rahman, Md. Asif Shahjalal, Md. Osman Ali, ByungDeok Chung, Yeong Min Jang · 2022
Low-resolution analog-to-digital converters (ADCs) in massive multiple-input multiple-output (m-MIMO) systems offer significant throughput increase for wireless communication. The low-resolution ADC part limits the performance of the transceiver, especially estimating the channels from highly quantized measurements. Mapping from quantized received signals to channel is very challenging. Since the channel response has the characteristic of sparsity, we have leveraged a fast and flexible denoising convolutional neural network (FFDNet) based channel estimation scheme for the m-MIMO system with one-bit ADCs treating the channel matrix as a 2D natural image in this letter. FFDNet can effectively learn from sufficiently large training datasets and is exploited to estimate the channel in the m-MIMO system equipped with one-bit ADCs. We have investigated the exhibited performance of the FFDNet based channel estimation through simulation results. A significantly reduced normalized mean square error has been achieved in our proposed scheme.