Near optimal iterative channel estimation for KSP-OFDM

In this paper, we propose an iterative 'turbo' channel estimation algorithm for known symbol padding (KSP) OFDM, where the guard interval is filled with pilot symbols. Additional pilot symbols are transmitted on some of the OFDM carriers. The channel estimation algorithm is based on the Expectation-Maximization (EM) algorithm. In the initialization phase of this iterative algorithm, the received time-domain samples are first converted to the frequency domain, and the initial channel estimate is based on the observation of the pilot carriers only. Then the EM-algorithm switches back to the time-domain and updates the channel estimates until convergence is reached. The proposed estimator performs very good: the MSE performance of the proposed estimator is close to the Cramer Rao lower bound (CRB) corresponding to the all pilots case, for the SNR region of practical interest, and the resulting bit error rate essentially coincides with the case of the perfectly known channel