Specialized non-smooth optimization techniques for controller design and tuning

Thumbnail

Event details

Date 23.05.2014
Hour 10:15
Speaker Prof. Pierre Apkarian, ONERA and Université Paul Sabatier
Location
Category Conferences - Seminars
This seminar will give an overview of recent non-smooth optimization techniques for linear controller synthesis and tuning.
Non-smooth techniques have practical advantages:
  • They can be used to optimize arbitrary control architectures made of single- or multiple-loop control arrangements from a collection of structured control elements including PIDs, reduced-order state-space models, transfer functions, observers, as well as custom controller blocks.
  • They can handle multiple design requirements hard or soft such as  H2, H-infinity, margins, pole constraints, …
  • They naturally lend themselves to multiple models as a way to enhance robustness.
These techniques however come with a local certificate as dictated by NP-hardness of the problems.
They can serve as good benchmarking tools for other approaches and heuristics.

Beyond highly specialized non-smooth algorithms, a brief account of the routines HINFSTRUCT and SYSTUNE from the
MATLAB Robust Control Toolbox will be presented.

Bio: Pierre Apkarian received the Ph.D. degree in control engineering from the Ecole Nationale Supérieure de l'Aéronautique et de l'Espace (ENSAE),  France,  in 1988. He was qualified as a Professor from Université Paul Sabatier (UPS) in both control engineering and applied mathematics in 1999 and 2001, respectively. Since 1988, he has been a Research Scientist at ONERA (Office National d'Etudes et de Recherches Aérospatiales)  and an Associate Professor at Institut de Mathématiques, UPS.  His research interests include robust and gain-scheduling control theory and LMI methods.  More recently, his research has focused on the development of specialized non-smooth programming techniques for control system design. Pierre Apkarian has served as an associate editor for the IEEE Transactions on Automatic Control and contributed to the non-smooth synthesis tools in the  Robust Control Toolbox of MATLAB.