On reduced-complexity iterative channel tracking in turbo coded OFDM

This paper proposes a receiver that combines iterative channel updates with iterative turbo decoding for high mobility OFDM systems. We particularly focus on reducing the number of the channel update iterations and the complexity of each iteration. A low-complexity time-frequency channel estimator is proposed that tracks the channel variation within one code window and predicts the channel for the next window. The channel estimator can use simple hard decisions or more computationally intensive "soft" decisions, in which case the estimator is derived from an expectation-maximization (EM) principle. Simulation results are provided under a varying number of channel updates that use hard or soft decisions. As an example, a very simple hard decision based receiver for a rate-1/2 turbo coded 16 QAM OFDM system performs within 1 dB of a pilot-based receiver at 500 Hz Doppler, after three channel iterations that are interlaced with eight max-log-MAP decoding iterations. The proposed techniques can significantly reduce pilot symbol overhead with minimal performance degradation.