Estimators for Time-Variant Channels Applied to UMTS-HSDPA

In this work we analyze the mean-square error (MSE) of two correlative channel estimators in time-variant channels. Both estimators make use of orthogonal basis functions that approximate the time-variant channel. The first approximation is based on a set of time shifted rectangular basis (RB) functions. The second approximation makes use of Slepian basis (SB) functions. We calculate the number of basis functions that minimize the MSE for the RB-and the SB-approach. Both estimators are designed for the full velocity range using the maximum velocity as design parameter only. Further, we apply the two estimators to the high-speed downlink packet access (HSDPA) mode of UMTS and illustrate the impact of the MSE on the throughput. With the SB-approach we observe a gain of roughly 0.4 dB over the RB-functions in case of an i.i.d. Rayleigh multipath channel with five taps