|
| Previous Up Next |
Identification of non-linear systems: identifiability and experiment design |
|
|
| Keywords: |
System identification and estimation, Modeling and analysis, Nonlinear and LPV systems
|
|
System identification is the scientific exercise that consists of determining a mathematical model of a real-life process (the true system) based on input-output data. A very important identification method is the ``prediction error identification’’ method. In prediction error identification, models can be identified based on data collected both in open loop and in closed loop. Based on these data, the model can be then determined within a given model structure by minimizing a least square criterion. The research in this project deals with prediction-error methods for the identification of non-linear systems. Fundamental research questions are:
- What are the neccessary and sufficient conditions (such as conditions on the richness of the input signal) that lead to consistent identification of a particular nonlinear model structure? Potential nonlinear model structures that will be investigated includes the linear parameter varying (LPV) model structure.
- What is the optimal experiment design for a particular nonlinear model structure (such as the LPV model structure)?
|
| Previous Up Next |
|
