Robust THP Transceiver Optimization under Imperfect CSI with Spatial Correlation

This paper presents a robust Multipleinput multiple-output (MIMO) transceiver optimization method based on Tomlinson-Harashima precoding (THP) structure, assuming that statistical imperfect Channel state information (CSI) model with spatial correlation. The design process expresses the Minimum square error (MSE) in terms of only the precoding matrix, and then optimizes the precoding matrix itself. For three scenarios with different spatial correlation information, the lower bound of the MSE is minimized, and the precoding matrix achieving the minimum can actually be chosen such that the lower bound is tight. The robustness and effectiveness of the proposed design is validated by simulations.