Filtering and Estimation Methods for Industrial Systems

Filtering and stochastic estimation methods are proposed for the control of linear and nonlinear dynamical systems. Starting from the theory of linear state observers the chapter proceeds to the standard Kalman filter and its generalization to the nonlinear case which is the Extended Kalman Filter. Additionally, Sigma-Point Kalman Filters are proposed as an improved nonlinear state estimation approach. Finally, to circumvent the restrictive assumption of Gaussian noise used in Kalman filter and its variants, the Particle Filter is proposed. Applications of filtering and estimation methods to industrial systems control with a reduced number of sensors are presented.