SMP model for movement-based registration with general cell sojourn time

This paper presents an exact analytical model and performance evaluation for movement-based registration (MBR). Firstly, we derive an exact model that considers the sojourn time at each state by using a semi-Markov process (SMP). Secondly, unlike previous studies we separate the implicit registrations from real location registrations in the state definition for the SMP. Lastly, we classify a ring into several groups of cells if the stochastic property of a group of cells is different from those of the other groups of cells in the same ring. Using our new model, the numerical results are obtained to compare with the previous model. It is shown that MBR should be evaluated by our new model to get an accurate performance especially when generally distributed cell sojourn time is considered.