Comparison of Low Complexity OFDM Channel Estimation Techniques

We address pilot-symbol aided channel estimation (PACE) for OFDM. To reduce the complexity of the optimum minimum mean squared error (MMSE) estimator, two sub-optimum estimators are investigated. First, an estimator utilizing an FIR filter of reduced length. The filter coefficients are determined according to a Wiener filter with model mismatch, i.e. the channel conditions which are assumed to compute the filter weights are chosen according to a typical scenario, and may be different from the actual channel. Second, channel estimation based on the discrete Fourier transform (DFT) is considered. By utilizing the fast Fourier transform (FFT) algorithm, interpolation, which is an integral part of PACE, can be performed very efficiently. However, the DFT-based estimator is very sensitive to the chosen channel model. The performance significantly degrades if the channel is non-sample spaced, i.e. the channel tap delays are not multiples of the sampling duration.