Modified channel estimation algorithms for OFDM systems with reduced complexity

Channel estimation techniques for Orthogonal Frequency Division Multiplexing (OFDM) systems, i.e., maximum likelihood (ML) and linear minimum mean squared error (MMSE) estimators are investigated in this paper. Although accurate estimation of a channel of known length (i.e. channel impulse response duration) can be obtained with conventional estimators, very few studies have been focused on the simplification of the ML and MMSE algorithms. Conventional channel estimators preset a fixed length for the channel. The knowledge of the channel length would, in principle, reduce the complexity of the ML and MMSE estimators. This is of particular interests for OFDM systems, due to the excessive matrix manipulation involved, when the channel duration and training sequence are very long. Hence, the complexity of channel estimation is of crucial importance, especially for time-varying channels, where it has to be performed periodically or even continuously. In order to simplify the OFDM channel estimator, a channel length estimation technique based on second order statistics of the received signal is proposed. No preamble and preliminary channel estimate are needed by the proposed technique. Analysis and simulation results will show that the channel length estimator is very robust to frequency offset.