Adaptive channel estimation for WCDMA RAKE receiver

Channel estimation in RAKE receivers is a well known problem. The presence of Doppler spread, and high, noise level signals will corrupt the data signal leading to unreliable demodulation. We propose the normalized least mean square (NLMS) prediction filter operating in a decision-aided mode. Linear interpolation based channel estimates are used to supply the reference signal for updating the filter coefficients. In addition, a signal-to-noise ratio (SNR) estimator will select the data estimate solely based on linear interpolation for signals during a deep fade. The proposed scheme is investigated in an FDD WCDMA system. Simulation results show that the proposed scheme can outperform conventional linear interpolation and decision-directed LMS prediction techniques under UMTS test scenarios. In particular, it will support robust estimation in low SNR regions, and will constrain the estimation error to within 0.9 dB loss from the ideal channel estimation benchmark, for a fast fading environment at 300 kmph.