Adaptive generative radio channel models

Generative channel models based on finite state models are introduced. The main method applied is separated quantisation of the various fading processes and their superposition to the channel process. A homogeneous finite state Markov model is proposed with constant transition probabilities to characterize the process. The transition probabilities are derived by two different algorithms. The fading processes are modelled separately. Thus, the transition probabilities are independent from parameters like e.g. the mean of the process. Due to this type of independence, these models are very useful for stochastic simulation and mathematical analysis of mobile radio networks. The resulting error process can be modelled by a discrete-time equivalent of a well a known process of queueing theory, the Markov modulated Poisson process. Some properties of the resulting error process are derived, which are very useful for evaluating the performance of mobile radio systems