Linear interpolation in pilot symbol assisted channel estimation for OFDM

In this paper, we investigate several efficient interpolation techniques for pilot symbol assisted channel estimation in OFDM. The interpolation methods studied include two dimensional (2-D) separable lowpass sinc interpolator with Kaiser window, 2-D separable Deslauriers-Dubuc (DD) interpolation and 2D discrete Fourier transform (DFT) based lowpass interpolation. The performances of these interpolators are compared with those of the well known minimum mean-square error (MMSE) 2-D separable Wiener filter and the perfect channel state information. It is shown that the Kaiser window based interpolator and DD interpolation are simple, robust, and outperform the 2-D DFT based lowpass interpolation as well as several existing interpolation methods proposed in the literature. These two schemes are suitable candidates for use in I-D and 2-D channel estimation.