The application of a generalised sliding FFT algorithm to prediction for a RAKE receiver system operating over mobile channels

Code division multiple access (CDMA) spread spectrum systems that employ RAKE receivers for data detection require an estimate of the channel impulse response to be made prior to extracting the data. In a rapidly fading environment, such as that experienced by mobile cellular systems, the application of a prediction algorithm to the problem of estimating the channel impulse response can improve the system performance over a non-predictive one. The paper describes the application of the least squares (LS) prediction algorithm to a RAKE receiver system, and present results obtained through Monte-Carlo simulation. The LS algorithm is computed using a computationally efficient sliding FFT based approach based on radix 4 butterflies. A novel prediction feedback method, called conditional decision feedback, for reducing the effect of decision feedback errors is presented. A typical urban environment was selected to evaluate the system performance, the COST 207 channel model for this environment being used to drive the simulation system.