Data-Driven Joint Demodulation and Decoding in THz Communication Systems

This paper explores the impact of impairments such as phase noise (PN), additive white gaussian noise (AWGN), and in-phase and quadrature (IQ) imbalance distortions in corrupting transmitted signals across a point-to-point terahertz (THz) communication system. Deep learning (DL) is leveraged at the receiver end to develop a deep neural network (DNN) based multi-label classification (MLC) that performs t he f unctions of denoising, demodulating, and decoding (DeNMC) in a single operational block. Datasets for the DL assisted DeNMC were trained offline over a wide range of signal-to-noise ratios (SNRs) and deployed online for real-time data processing at the receiver. In this paper, we show that artificial intelligence (AI) can help mitigate the PN effect and IQ imbalances. Simulation results proved that implementing DL assisted DeNMC yields better performance than the conventional non-AI receiver while achieving lower computational complexity.