| |
Identification and Control of a Nonlinear Mechatronic System
| [View]
[Download]
|
| Author(s) |
C. Hintz; B. Angerer; D. Schroeder |
| Abstract |
In this paper an identification algorithm and a nonlinear control strategy for a mechatronic
system is presented. In order to achieve accurate control results all linear parameters (e.g. iner-
tia) and nonlinear characteristics (e.g. friction) have to be known. To illustrate the theoretical
investigations a throttle is identified and controlled.
In the first part of this paper the identification algorithm is presented. The throttle is hereby
modelled as a two mass system (TMS) [8]. All linear parameters (inertias of the DC-Motor and
the throttle, spring- and damping-constant of the mechanical connection including the gear)
and nonlinear characteristics (friction and reset spring) are identified by a structured recurrent
neural network (SRNN).
Based on the identification results an adaptive PI-State-Space-Controller with a nonlinear
observer is developed. This PI-State-Space-Controller is compared with a state of the art PID-
Single-Loop-Controller regarding control accuracy, control response time and noise behavior.
Measurement results show the capabilities of the presented identification and control concept. |
| Download |
| Filename: | EPE-PEMC2002 - T10-001 - Hintz.pdf |
| Filesize: | 644.3 KB |
|
| Type |
Members Only |
| Date |
Last modified 2004-05-13 by System |
| |
 |