MASTER Topic 2014-2015 Adaptive control for ultra-fast - GIPSA-lab

MASTER Topic 2014-2015
Adaptive control for ultra-fast nanopositioning
Duration: 5 months,
Place : Control Systems Department, Gipsa-lab
Supervisors :
Alina VODA,
[email protected],
+33476826233
Gildas BESANCON, [email protected], +33476826230
Lukasz RYBA,
lukasz.ryba@@gipsa-lab.grenoble-inp.fr
Description
The proposed topic is about high-performance control of a nanopositioning
device, to be implemented on the so-called Nano-control experimental platform
developed in the Control Systems Department of Gipsa-lab (see picture below). This
system aims at ultra-fast nanopositioning of a flexible micro-cantilever, which is
illustrative of many applications in nano-sciences and nano-technologies.
Difficult aspects to be faced in such a system (when a high performance is
looked for), are related to its very fast dynamics, and consequently resonances,
uncertainties, or nonlinearities which are excited, in addition to noises. Our research
group already obtained very promising results for the positioning with an accuracy up
to the nanometer (or even Angström), and with a bandwidth around 1 kHz, by using
robust control methodologies (see references hereafter).
The challenge of the present project is the positioning of the cantilever over a
large interval of flexibility, going from its fixed point (with infinite rigidity) up to its most
mobile part (with rigidity = 0.01 N/m). In this context, the cantilever model can be seen
with a varying damping coefficient, from a damped model to a resonant one, and the
purpose is then to achieve the fastest nanopositioning which can be reached.
It is important to test first the limits of robust control, since such an approach
can remedy to moderate parameter variations as well as neglected dynamics (high
frequency).
The main part of the project is the design of an adaptive robust control that
should tackle the large variation of rigidity, and achieve a high bandwidth. An
interesting strategy can be multi-model adaptation, based on closed-loop identification.
The developed approaches will be first tested in simulation (Matlab/Simulink),
on an illustrative model, before being experimentally validated.
References :
L. Ryba, A. Voda and G. Besançon: “An LQG/LTR approach towards piezoactuator
vibration reduction with observer-based hysteresis compensation”, Proceedings
of 19th IFAC World Congress, Capetown, pp.5623-28, 2014
S. Blanvillain, A. Voda, G. Besançon, and G. Buche: “Subnanometer positioning and
drift compensation with tunneling current,” IEEE Transactions on Control Systems
Technology, 22(2):180–189, 2013.
I. Ahmad, A. Voda, G. Besançon, G. Buche: "Robust digital control approach for
high performance tunneling current measurement system", Control Engineering
Practice, Volume 20, Issue 7, pp. 643-653, Elsevier, 2012.
A. Voda (Ed.), “Micro and nanosystems and systems on chip – Modeling, control
and estimation “,Wiley, ISTE, Londres et New Jersey, 304 pages, 2010.
I.D.Landau « From robust control to adaptive control » Control Eng.Practice, vol
7,no10, pp1113-1124, 1999.
Narendra K.S., Balakrishnan « Adaptive control using multiple models », IEEE Tr.
on Aut. Control, AC-42, pp. 171-187, 1997.