.png){kind=icon}
.png){kind=icon}
Return
Iterative control decoupling tuning for precision MIMO motion systems: A matrix update method
.png){kind=icon}
PRE
AI
DOI：10.1016/j.isatra.2024.10.003.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Control decoupling is the basic control step for precision MIMO (Multi-Input-Multi-Output) motion systems. After decoupling, the MIMO logical controlled plant can be diagonal dominant so that the control design for each DOF (Degree of Freedom) can be separately treated. However, due to the manufacturing tolerances and the assembling errors, the actual CoG (Center of Gravity) and actuator positions are inconsistent with the designed values. The nominal static control decoupling matrix is inaccurate, which significantly deteriorates the decoupling performance. To address the problem, this paper develops a matrix update based iterative control decoupling tuning method. Tuning with feedback control signals to achieve accurate tuning is its first distinct feature. By employing rigid-body model information to construct more accurate basis functions, fast tuning can also be achieved, especially for relatively low control bandwidth occasions. Differing from the feedforward compensation based iterative control decoupling tuning method, the matrix update method improves the dynamics of the logical controlled plant, does not increase the control complexity and is more robust to the inaccuracy of the model information. Experimental results on the short-stroke module of a precision motion stage which is the key subsystem of the lithographic projection lens testing system present significant control decoupling performance improvement (for example the peak value of the Rx-DOF servo error caused by the ZDOF movement is reduced from 1.29 x10- 5 m to 3.19 x10- 6 m) after just two trials.
Keywords:
Control decoupling
Iterative tuning
Iterative learning control
MIMO motion systems