2005 Fiscal Year Final Research Report Summary
A Study on System Identification based on Input-Output Data with Uncertainty
| Project/Area Number |
16560378
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Control engineering
|
|---|
| Research Institution | Utsunomiya University |
|---|
Principal Investigator |
ADACHI Shuichi Utsunomiya University, Faculty of Engineering, Professor, 工学部, 教授 (40222624)
|
|---|
| Project Period (FY) |
2004 – 2005
|
| Keywords | system identification / non-liner system / optimization with constraints / unscented Kalman filter / hybrid identification |
|---|
| Research Abstract |
In this research, system identification problem in which input and output data are measured with uncertainty such as nonlinearity are studied. Three kinds of identification methods are considered:
(1) A system identification in the presence of nonlinear sensors : A novel system identification method which estimates sensor errors due to its nonlinearities, as well as system parameters, is proposed. The key idea is to formulate the identification problem as a constrained optimization problem. Then the system parameters are estimated by using quasi-Newton method.
(2) A hybrid modeling of piece-wise linear system : A driver model which can evaluate performance of obstacle avoidance assistance system algorithm is developed. The driving data s collected by small-scale traffic stream simulator. A hybrid modeling is applied to build the driver's avoidance behavior.
(3) A nonlinear system identification method using a unscented Kalman filter : Through numerical simulations and experimental data such as data from engine control system of automobile, the effectiveness of the proposed system identification method is verified.
|
