Adaptively Regularized Least-Squares Estimator for Decision-Directed Channel Estimation in Transmit-Diversity OFDM Systems

Decision-directed channel estimation (DDCE) in orthogonal frequency-division multiplexing systems with transmit diversity can be simplified by cancelling inter-antenna interference to separate each channel estimation. Conventional methods involve using the standard least-squares (LS) estimator and exhibit severe decision error propagation. This study proposes an adaptively regularized LS estimator to improve the DDCE performance. The proposed method adopts the latest DDCE estimate as a priori information in the regularized LS estimator, and adaptively determines the regularization parameter according to the mean-squared error of the a priori estimate and that of the standard LS estimate. Simulation results demonstrate the considerable improvement of the proposed algorithm.