Low-complexity channel estimation for WCDMA random access

This paper investigates various low-complexity channel estimation techniques applied to the random access channel of wideband CDMA (WCDMA) specified by 3GPP for IMT-2000/UMTS. Good access performance is important for efficient packet data transmission. Different schemes using slot-based or symbol-based pilot-assisted channel weights are considered. Several interpolation methods and post-filters are compared. A unified frequency domain post-filtering for interference and noise suppression is employed. Link level simulation results for typical indoor and vehicular fading channels are presented and compared with performance of ideal (undistorted) channel weights. Performance criterion is the frame error rate (FER) vs. average signal-to-noise-ratio E_b/N₀. It turns out that symbol-based channel estimation can lead to significant improvements of 1-1O dB compared to conventional slot-based approaches without post-filter; especially for high speed scenarios (>200 km/h). A good choice for all scenarios is a non-causal forward-backward-multiplication-method (FBMM) filter employing modified polar interpolation