An adaptive estimator with learning for a plant containing semi-Markovian switching parameters

A nonlinear state estimator has been developed to solve the problem of a randomly switching plant operating in white Gaussian noise. A switching plant by definition has certain key parameters that can vary randomly within a finite set of real values. In modeling the stochastic system it will be assumed that the parameter variations will be described by a semi-Markov process. By incorporating the semi-Markovian concept into a Bayesian estimation scheme an adaptive state estimator was developed which could handle the switching plant or switching environment problem without computer storage increasing as time progresses.