Frequency-Domain Data-Aided Channel Estimation for KSP-OFDM

In this paper, we propose a new low-complexity data-aided channel estimation algorithm for known symbol padding (KSP) OFDM. Besides the pilot symbols in the guard interval, additional pilots are put on some of the OFDM carriers. The received time-domain samples are first converted to the frequency domain, and the channel estimation is based on the observation of the pilot carriers only. The proposed estimator is compared to the subset estimator from [1]. Although performing slightly worse than the subset estimator at low SNR, the proposed estimator outperforms the subset estimator at high SNR, as the former does not suffer from an error floor in the MSE. The MSE resulting from the proposed estimator is inversely proportional to the number of pilot carriers and is independent of the FFT size, the pilot carrier positions and the used pilot sequence.