Time Domain LMMSE Channel Estimator Based on Sliding Window for OFDM Systems in High Mobility Situations

Abstract

Recently, Wideband Orthogonal Frequency Division Multiplexing (OFDM) based systems like Long Term Evolution (LTE) have been proposed for future railway communication systems. In such high speed situations, Doppler phenomena, related to the terminal's high speed, affects the use of OFDM systems by producing a loss of orthogonality between different carriers. Thus, the channel estimation and equalisation become a challenging problem especially when the number of pilot carriers is limited. This paper investigates the use of time domain channel estimation for OFDM systems in high speed scenarios, and for a sparse pilot distribution. A Time Domain Linear Minimum Mean Square Error (TD-LMMSE) filter is designed for channel estimation. The TD-LMMSE emulates the effect of a sliding window, and takes into consideration the interference caused by the data field. The expression of the Mean Square Error (MSE) of linear estimators is also developed. Furthermore, a simple technique for the estimation of the 2nd order channel statistics is proposed. Simulation results, showing a significant gain in terms of MSE and Bit Error Rate (BER) compared to other state of the art algorithms are reported over a wide Signal-to-Noise Ratio (SNR) range.