Adaptive Long-Range Prediction of Three-State LMS Channel Model

Abstract

In order to provide the accurate real-time knowledge of future channel state information for the adaptive transmission of narrowband land mobile satellite (LMS) communication systems, an adaptive long-range prediction (ALRP) method, suitable for a three-state LMS channel model at S-band, is proposed. The very slow shadowing conditions' variations of satellite propagation channel can be described as a three-state Gilbert---Elliot channel model based on Markov chain. The channel state of future long-range is predicted based on weighting prediction while the parameters of linear auto-regression model are updated using an iterative adaptive tracking algorithm. The future channel fading series are predicted by the latest observations within current channel state. Simulation results show that the proposed method can be used to predict the future long-range channel state and fading series of three-state Gilbert---Elliot channel model accurately; and it is observed from the mean square error results that the prediction performance of the ALRP method is much more accurate than that of long-range prediction method. Moreover, the method has advantage of real-time and low-complexity which will make the adaptive transmission techniques feasible for LMS channel models.