Doubly Selective Channel Estimation for OFDM Systems

In order to estimate doubly selective channels in orthogonal frequency division multiplexing (OFDM) systems, we derive a linear minimum mean square error (LMMSE) channel estimator by considering fast time-variant channel estimation by Wiener filtering. To reduce the computational complexity of the LMMSE estimator, we also present an approximate LMMSE (ALMMSE) estimator. This estimator obtains partial channel information by exploiting LMMSE estimation and finds time-variant channel impulse responses by interpolating the channel information with the discrete orthonormal Legendre polynomials. Furthermore, we propose an iterative LMMSE estimation scheme to significantly improve channel estimation performance. By exploiting both detected information symbols and given pilot symbols as new pilot symbols, this scheme iteratively uses the proposed approximate LMMSE estimator to update channel estimates. Simulation results demonstrate the effectiveness and good performance of the proposed channel estimation schemes