In this paper, we investigate the tracking of time-varying channels in OFDM systems using both preamble and pilot symbols. A regularized least-squares method will be used to incorporate the initial channel estimate from the preamble to improve the performance of pilot-assisted channel estimation.
Instead of using a fixed regularization parameter, the proposed method adaptively determines the regularization parameter according to mean-squared error of the latest channel estimate and that of the standard LS estimate. Simulation results demonstrate the proposed method can outperform the standard LS method with lesser pilots, thus achieving the reduction of pilot overhead especially at lower SNR environments.