Accurate Wideband Channel Estimation for THz Massive MIMO Systems

This letter addresses the wideband channel estimation problem for terahertz (THz) massive multiple input multiple output (MIMO) systems. The proposed wideband channel estimation scheme consists of three stages. First, the wideband channel estimation is formulated as a joint sparse signal recovery problem based on an on-grid signal model; and the coarse estimates of angles of arrivals (AoAs) and angles of departure (AoDs) of the paths can be obtained by an orthogonal matching pursuit (OMP) based method. Then, an expectation-maximization (EM) based method is proposed to refine the estimates of AoAs and AoDs by recasting the wideband channel estimation problem. Finally, given the estimates of AoAs, AoDs and composite path gains, the delays and gains of the paths are further extracted with the help of constructed dictionary matrices. Simulation results show that the proposed scheme can improve channel estimation accuracy by up to 5dB of the signal-to-noise ratio (SNR) compared to the perfect AoAs and AoDs aided oracle least square (LS) method.