A comparative study of iterative channel estimators for mobile OFDM systems

We investigate two iterative channel estimators for mobile orthogonal-frequency division multiplexing. The first estimator is based on iterative filtering and decoding whereas the second one uses an a posteriori probability (APP) algorithm. The first method consists of two cascaded one-dimensional Wiener filters, which interpolate the unknown time-varying two-dimensional frequency response in between the known pilot symbols. As shown, the performance can be increased by feeding back the likelihood values at the output of the APP-decoder to iteratively compute an improved estimate of the channel frequency response. The second method applies two APP estimators, one for the frequency and the other one for the time direction. The two estimators are embedded in an iterative loop similar to the turbo decoding principle. As shown in detail, this iterative estimator is superior and its performance is independent of whether the chosen time-frequency pilot grid satisfies the two-dimensional sampling theorem or not. The bit-error rate as a function of the signal-to-noise ratio is used as a performance measure. In addition, the convergence of the iterative decoding loop is studied with the extrinsic information transfer chart.