A context-tree based model for quantized fading

Fading is modeled by applying Contest Tree Pruning (CTP) to the sequence of fading states that results from quantizing a sequence of real-valued fadin g measurements. There results a family of finite state Markov chains of inc reasing order that represent the channel with increasing fidelity. A fading state may be the union of chain states, A particular model mag be chosen f rom that family according to an application-specific criterion. The resulti ng models facilitate analysis and simulation of wireless systems that explo it channel memory. An example criterion compares each model's fading state sojourn time statistics with those of the original sequence.