Tom Oomen

Research on Inferential Motion Control and Extensions

Recently, our inferential control paper has been accepted for journal publication! See also the extended report. The idea for this paper already started about a decade ago in the beyond-rigid-body control discussions, where flexible modes have an important role, as is also argued in Oomen et al., 2014.

The basic idea is that in traditional rigid-body motion systems, geometric relations exist between the sensor measurements and the point of interest. In next-generation systems, internal deformations of the system lead to a dynamical relation between measured variable and performance variable at the point of interest. This is shown in the following figures for a lithographic process and a printer.

The basic idea is that model-based controllers are observer-based. Thus, if we would use a good model, then the controller will be able to internally estimate the performance variable and achieve good performance!

A strong focus in the theoretical developments has been on incorporating model uncertainty in both system identification and robust control. Model uncertainty is of vital importance here, since the model is used for prediction. Important developments that extend the results in Oomen and Bosgra, 2012, are presented in

The theoretical developments have been applied on a protoype motion system where the performance variable is not directly measured in the feedback loop.

The results are documented in

The above results have been demonstrated on a prototype motion system. A full model-based control design requires an extended weighting filter selection, e.g., based on H-infinity-optimization. To facilitate industrial implementation, also an approach is developed where the modal deformations are estimated using observers, e.g., based on Kalman gains. The approach and experimental application to our laboratory system and a wafer stage FEM model are provided in the following thesis.

Recently, practical relevance has been shown on a next-generation wafer stage. Interestingly, through a different approach, i.e., exploiting additional actuators and sensors, it has been shown that the inferential performance can also be drastically improved. This is reported in the following documents.

  • Experimental analysis of the influence of structural deformations on a new lightweight industrial motion system on its positioning accuracy using a systematic identification-and-robust-control design framework
    Joris Termaat
    M.Sc. Thesis, Eindhoven University of Technology, Eindhoven, The Netherlands, 2011

  • Exploiting additional actuators and sensors for nano-positioning robust motion control [preprint|link]
    Robbert van Herpen, Tom Oomen, Edward Kikken, Marc van de Wal, Wouter Aangenent, and Maarten Steinbuch
    IFAC Mechatronics, Invited paper, 24(6): 619-631, 2014

Ongoing work focusses is in two directions. In one direction, both theoretical and practical aspects are investigated where the performance variable is measurable in a batch-to-batch fashion, i.e., not in real-time. The basic idea is to use approaches such as Iterative Learning Control (ILC) to enhance “inferential performance”. Initial results are reported in:

  • Aspects in inferential iterative learning control: a 2D systems analysis
    Joost Bolder, Tom Oomen, and Maarten Steinbuch
    In Proceedings of the 52nd IEEE Conference on Decision and Control, Los Angeles, California, United States, 2014

  • On inferential iterative learning control: with example on a printing system [pdf]
    Joost Bolder, Tom Oomen, and Maarten Steinbuch
    In Proceedings of the 2014 American Control Conference, 1827-1832, Portland, Oregon, United States, 2014

In another direction, we are investigated inferential performance improvement in thermal control. An specific application related to lithography is described in

  • Positioning Performance Enhancement via Identification and Control of Thermal Dynamics: A MIMO Wafer Table Case Study
    Juan Guo
    M.Sc. Thesis, Eindhoven University of Technology, Eindhoven, The Netherlands, 2014

The success of all the above work is due to the hard and excellent work of many people involved in this research, including

  • Active researchers at TU/e-ME-CST: Joost Bolder and Robbert Voorhoeve

  • Previous reseachers at TU/e-ME-CST: Okko Bosgra, Erik Grassens, Juan Guo, Ferdinand Hendriks, Robbert van Herpen, Joris Termaat

  • Industrial collaborators from ASML, Oce, Philips: Wouter Aangenent, Jeroen van Helvoort, Sjirk Koekebakker, Marc van de Wal
    and many others

Note that all figures shown on this page can be found in the mentioned papers. Please follow the guidelines regarding copyright and references when citing these.