Kalman filter estimator for WCDMA FDD Rake receiver

Wideband demodulation performance can be hindered by the channel estimator. In particular, the fading property of the channel and high noise level signals will introduce a degree of discrepancy in the channel estimate. In this paper, we investigate the Rake receiver performance incorporating decision-aided Kalman filter estimation. The Kalman predictor estimates the present channel value, based on a first-order Gauss Markov model, for subsequent maximal ratio combining. Moreover, the decision-aided component uses data estimates based on linear interpolation, to correct the initial predicted complex channel amplitude. Simulation results show that the proposed scheme can overcome the channel fading effect for moderate Doppler spreads, and can offer significant improvement in performance compared to conventional schemes, in terms of estimation and complexity.