Abstract
In this paper we present a framework for identification of hybrid systems, based on switched input/output (SIO) models. The property of linearity for SIO models is defined. We further define hybrid generalizations of classic linear input/output models, such as switched ARX, switched ARMAX, and switched output error models which are shown to be linear in the SIO sense, and Box-Jenkins model, which is not linear. A Bayesian algorithm for identification of linear SIO models is developed. Operation of the developed algorithm is demonstrated on an example.
| Original language | English |
|---|---|
| Title of host publication | Proceeedings of the 17th International Symposium on Mathematical Theory of Networks and Systems (MTNS 2006) July 24-28, 2006, Kyoto, Japan |
| Publisher | MTNS |
| Pages | WeA05.5- |
| Publication status | Published - 2006 |
| Event | 17th International Symposium on Mathematical Theory of Networks and Systems (MTNS 2006) - Kyoto, Japan Duration: 24 Jul 2006 → 28 Jul 2006 Conference number: 17th |
Conference
| Conference | 17th International Symposium on Mathematical Theory of Networks and Systems (MTNS 2006) |
|---|---|
| Abbreviated title | MTNS 2006 |
| Country/Territory | Japan |
| City | Kyoto |
| Period | 24/07/06 → 28/07/06 |