Tensor-Based Joint Symbol Detection, User Localization and Tracking for THz Systems with Dual-Wideband Effects

In this paper, we develop a tensor-based joint symbol detection and user sensing scheme for the three-dimensional uplink terahertz system with dual-wideband effects. To leverage the low-rank decomposition of tensor, the received matrix-form pilot at the selected subcarrier is rearranged into a real-valued tensor-form one, and the corresponding algorithm is designed for angle of arrival and Doppler estimation. Subsequently, symbol detection and time delay estimation are conducted by utilizing the received signals at the remaining subcarriers. Finally, user localization and tracking are achieved by using the estimated sensing parameters along with the location variations between adjacent time instances. Simulation results show that the proposed scheme provides superior joint symbol detection and user sensing performance compared with the existing complex-valued matrix-based and tensor-based methods.