Wavelet-based massive MIMO-NOMA with advanced channel estimation and detection powered by deep learning

Abstract

This paper presents a massive multiple-input multiple-output (mMIMO) non-orthogonal multiple access (NOMA) system’s channel estimation and detection method which uses a deep learning algorithm to address the issue of erroneous signal detection caused by imperfect channel state information (CSI), user interference and channel noise. The proposed network applies deep learning (DL) model to a discrete wavelet transform (DWT)-based orthogonal frequency-division multiplexing (OFDM) approach, aiming to reduce noise and inter-channel interference (ICI) while maintaining compatibility with the existing networks. A significant advantage of this network is the reduction of transmission overhead as it can estimate and detect symbols with and without pilot signals. The study shows that the DL-based channel estimation and detection method proposed in this paper enhances the successive interference cancellation (SIC) process compared to the conventional linear scheme-based SIC in the mMIMO-NOMA network. Additionally, the wavelet-based mMIMO-NOMA is compared to the traditional fast-Fourier-transform (FFT) based mMIMO-NOMA in terms of symbol error rate (SER) for both perfect and imperfect CSI. The simulation results show that the proposed method outperforms the traditional method.