An adaptive long-range prediction based on two-state LMS channel model at S-band

Abstract:

Considering the narrowband two-state land mobile satellite channel model with variable model parameter at S-Band, an adaptive long-range prediction method is proposed based on weighting prediction. Firstly, a two-state Markov Gilbert-Elliot channel model with an ability of describing shadowing conditions of satellite communication downlink is established. And then, the future long-range channel state is predicted by weighting prediction, and the coefficients of linear auto-regression model are updated by iterative adaptive tracking method using the least mean square algorithm. Finally, the future channel fading series are predicted. Simulation results show that the proposed method not only can be used to predict the future long-range channel states and fading series accurately, but also improve prediction performance compared with the long-range prediction method. Moreover, this method has ability of real-time and low-complexity and can be used in the adaptive transmission performance analysis of narrowband LMS communication systems.