Multiple-Model Approach to Hybrid Estimation
by
X. Rong Li
Department of Electrical Engineering
University of New Orleans
New Orleans, LA 70148, USA
xli@uno.edu, Tel: 504-280-7416, Fax: 504-280-3950
Abstract
Hybrid estimation refers to estimation of a hybrid random process with jump as
well as diffusion components, such as state estimation for a hybrid system,
which includes both continuous- and discrete-valued state variables. It unifies
conventional estimation and decision, and provides a powerful framework for
many inference problems (e.g., detection, estimation, recognition, and identification)
involving both structural and parametric uncertainties/changes, among others.
An introduction to hybrid estimation is given.
One of the most natural and popular approach to hybrid estimation is the
so-called multiple-model method. It uses more than one model at every time to
represent or cover possibilities of the unknown structures; the overall
estimate is a combination of single-model-based estimates; and the
probability/likelihood of a model provides a quantitative measure of the
unknown true structure.
An overview of multiple-model estimation approach is presented. Its basic ideas
are explained. Its three generations (autonomous, interacting, and variable
structure) are described. An application example for detection and
identification of aircraft sensor and actuator faults are given. Historical
milestones, recent advances, and open problems of the multiple-model approach
are addressed.
Notes from first lecture
Close
|