Optimized Iterative Channel Estimation for OFDM

In this paper(1), we investigate iterative channel estimation for coded OFDM over frequency-selective, time-varying channels. Provided that an initial channel estimate is available, iterative channel estimation (ICE) utilizes data symbol estimates fed back by, the channel decoder. The efficiency of these iterative approaches, however depends on the reliability of the data estimates. In the presence of decision feedback errors, the channel estimation performance is severely degraded. We propose an optimized ICE based on the minimum mean squared error (MMSE) criterion that incorporates statistical information of decision feedback errors, readily available in the form of soft information. Moreover, an efficient implementation of the channel estimation unit is described. Simulation results indicate that the new scheme outperforms existing techniques in terms of channel mean-squared error (MSE) and bit error rate (BER).