Robust 2D+1D adaptive channel estimation scheme in OFDM systems

Linear pilot-aided channel estimation in orthogonal-frequency division multiplexing (OFDM) systems is carried out optimally via 2D Wiener filters. The standard approach employed in conventional receivers is based on cascaded 2x1D Wiener filtering, which has a lower complexity but also suboptimal performance. The 2D+1D filtering scheme proposed in this paper combines the benefits of the two other schemes and is not affected by some restrictions of 2x1D filtering as shown by theoretical analysis of such schemes. For comparison, the mean-squared error (MSE) performance of the different channel estimation schemes is evaluated. Based on this analysis, results from full link-level simulations w.r.t. the digital video broadcasting standard for satellite services to handheld devices (DVB-SH) are presented applying adaptive channel estimation algorithms instead of the Wiener filters. The novel 2D+1D scheme shows enhanced bit error rate (BER) performance since it provides more accurate channel estimates in high Doppler channel conditions compared to the conventional cascaded 2x1D schemes.