On the Effectiveness of Various Machine Learning Algorithms for THz Channel Estimation

Terahertz communication is one of the most promising wireless communication technologies, due to its capability to provide high bitrates. THz frequencies suffer however from high signal attenuation and signal degradation, which makes the THz channel modeling and estimation fundamentally hard. On the other hand, channel estimation of THz transmission system is critical for THz systems to be practically adopted in wireless communications. We consider the problem of channel modeling with deterministic channel propagation and the related physical characteristics of THz bands, and study the effectiveness of various machine learning algorithms to estimate the channel. We apply different machine learning algorithms for channel estimation, including neural networks (NN), logistic regression (LR), and projected gradient ascent (PGA). Numerical results show that PGA algorithm yields the most promising performance at SNR=0 dB with NMSE of -12.8 dB.