Two new approaches to channel prediction based on sinusoidal modelling

Prediction of communication channels is important, for example to enable adaptive transmission strategies. In this paper, two new predictors are proposed based on sinusoidal modelling. First a combined LMMSE (given frequency estimates) and LP is proposed. Then an improved LMMSE prediction is proposed by modelling the frequency estimate errors as Gaussian random variables. Both predictions have better performances than the LMMSE prediction (given frequency estimates) in computer simulations