A channel estimation method based on frequency-domain pilots and time-domain processing for OFDM systems

Abstract

A channel estimation method based on frequency-domain pilots and time-domain processing (FPTP) is proposed for OFDM systems in time-varying channels. Through discrete Fourier transform (DFT), the channel impulse response (ClR) is estimated by using the frequency-domain pilots, which is equivalent to the CIR estimated by cyclic correlation using time-domain pilot sequence. Hence, the channel estimation error, caused by the spectral leakage of the DFT, can be reduced by canceling the non-idea auto-correlation property of the pilot sequence in time-domain. At the same time, the complicated cyclic correlation computation in time-domain is evitable. The theoretical analysis and simulation results show that the proposed method outperforms the traditional DFT-based method and Wiener filter algorithm in time-varying multipath fading channels. Moreover, the uncoded bit error rate (BER) performance of proposed algorithm is very close to that with perfect channel knowledge, especially in fast fading channels.