Polynomial Prediction of Noise Shaping Rayleigh Fading

In this paper, polynomial prediction of Rayleigh fading power signal produced by noise shaping is investigated. We focus on the processing of the measured input signal for power control purposes of a Code Division Multiple Access (CDMA) mobile station receiver. Predictive filtering based on a polynomial signal model is proposed, and a class of FIR-type polynomial predictors is investigated with simulations. Two alternative schemes are examined; direct prediction of the squared power signal, and prediction of the in-phase and quadrature signal components separately. The simulations show that finite impulse response (FIR) polynomial predictors can provide smoothed signal power samples with the signal-to-noise ratio (SNR) improved by ca. 5 dB, without any delaying of the signal. The results show that polynomial prediction is a highly potential tool for delayless filtering of additive noise and smoothing of fast fading of the power signal.