Information Theoretic Approach to Finite State Markov Modelling of Time Varying Channels

Abstract

The time varying binary symmetric channel (TV-BSC) model is presented in this paper. The TV-BSC is developed to be the simplest model which captures the most important non-trivial attributes of general time-varying channels. An information theoretic analysis verifies that this model holds the essential elements of the behavior of time varying channels. The influence of the channel uncertainty is captured serves to significantly differentiate the TV-BSC model from other two-state Markov models, especially from the the Gilbert-Elliot model. This elementary form of time varying channel model has the information capacity that is calculated, using several methods. Given this capacity, the most efficient way for the time varying communications channel response estimation can be considered. Finally, accurate approximations for the TV-BSC information capacity are found, providing better inside into the capacity analysis of time varying communication channels.