Search

This work considers the problem of observer design for continuous-time systems with sampled output measurements (continuous-discrete time systems). In classical literature and in many applications, the continuous-discrete time extended Kalman filter (EKF) is used in order to tackle this problem. In this work, using a normal form characterizing the class of nonlinear uniformly observable single output nonlinear systems, it is shown that a particular stationary solution of a continuous discrete time Lyapunov equation can be used in order to design a constant high gain observer. Explicit conditions are given to ensure global convergence of the observer. Finally, an illustration of this result is given using an example of a biological process. [ABSTRACT FROM AUTHOR]

The design of high gain observers is usually based on normal forms of observability. If the system is observable for every input (uniform observability), the gain of the observer does not required a solution to differential equation. For multiple input multiple output (MIMO) non-uniformly observable systems, we give here a sufficient condition that the input must satisfy in order to design an observer. Unlike uniformly observable systems, the observer gain of non-uniformly observable systems is derived from a Lyapunov differential equation. [ABSTRACT FROM AUTHOR]

In this note, we present two system structures that characterize classes of multi-input multi-output uniformly observable systems. The first structure is decomposable into a linear and a nonlinear part while the second takes a more general form. It is shown that the second system structure, being more general, contains several system structures that are available in the literature. Two high gain observer design methodologies are presented for both structures and their distinct features are highlighted. [ABSTRACT FROM AUTHOR]