Moving Horizon State Estimation for Nonlinear Systems: Application to a Chemical Reactor CSTR

摘要

This paper considers the state estimation problem for the nonlinear systems. The approach, as it is based on the moving horizon technique, aims at studying the spirit of the estimator and its implementation simultaneously. This technique, in fact, allows for transposing the observation problem to an optimization one, it also consists in minimizing the gap between the measurement of the system and its prediction on a horizon of preset time. The optimization algorithm that is used is of Levenberg-Marquardt. The developed method, as a result, is validated on the nonlinear model of a continuous chemical reactor.