A Critical Discussion of the Continuous-Discrete Extended Kalman Filter
Systematic methods and tools for managing the complexity
Process Control (T4-8P)
Keywords: State Estimation, Extended Kalman Filter, Stochastic Differential Equation
In this paper, we derive and apply a novel numerically
robust and computationally efficient extended Kalman
filter for state estimation in nonlinear continuous-discrete
stochastic systems. The continuous-discrete extended Kalman
filter is applied to the Van der Vusse reaction example. This
example is a well-known benchmark for nonlinear predictive
control. Using the Van der Vusse example, we demonstrate
inherent limitations of the extended Kalman filter and sensor
structure for unbiased state estimation. In particular, we
demonstrate that the convergence rate of the concentration
estimate in the Van der Vusse system is limited by the frequency
of concentration measurements. These limitations limit the
achievable performance of any closed-loop system including
nonlinear model predictive control.
See the full pdf manuscript of the abstract.
Presented Tuesday 18, 13:30 to 15:00, in session Process Control (T4-8P).
