Detection and estimation for abruptly changing systems

The problem of state estimation and system structure detection for discrete stochastic dynamical systems with parameters which may switch among a finite set of values is considered. The switchings are modelled by a Markov chain with known transition probabilities. A brief survey and a unified treatment of the existing suboptimal algorithms are provided. The optimal algorithms require exponentially increasing memory and computations with time. Simulation results comparing the various suboptimal algorithms are presented.