Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.


Efficient Nonlinear Model Predicitve Control via Set Membership approximation techniques

Nonlinear Model Predictive Control (NMPC) has received an ever-increasing attention for industrial applications, due to its capability of treating different kinds of control problems in a quite general framework, in the presence of both linear and nonlinear system models, and to its efficiency in handling constraints on the input, state and output variables. However, NMPC can be effectively applied only if the employed sampling time is sufficiently large, so to allow the real-time solution of the underlying optimal control problem. On the other hand, the features of NMPC make this technique interesting also for systems with “fast” dynamics, which require high sampling frequencies that do not allow to solve the optimization problem in real-time. This issue motivates the significant research effort that has been devoted in recent years to develop techniques for the efficient implementation of NMPC. Moreover, in many applications (e.g.automotive), the capability to obtain good control performance with low-cost hardware is a point of great importance and a key for economical success: this aspect further motivates the research studies proposed in the literature to improve the efficiency of NMPC and to enable its implementation also on processors with limited computational performance. A viable approach is the use of function approximation techniques to derive off-line an approximated NMPC law, with lower on-line computational burden. In this talk, the use of Set Membership approximation methodologies to compute approximate NMPC laws will be presented, with particular attention to the analysis of the approximation error and of its effects on closed loop stability and performance, to the tradeoff between accuracy and complexity and to the approximation of discontinuous NMPC laws. Some numerical examples and an automotive case study will be employed to show the features of the described approach.

Type of Seminar:
Public Seminar
Dr. Lorenzo Fagiano
Politecnico di Torino, Italy
Jun 01, 2010   17:00

ETH Zurich, Gloriastrasse 35, Room ETZ J 91
Contact Person:

Prof. M. Morari
File Download:

Request a copy of this publication.
Biographical Sketch:
Lorenzo Fagiano received the Master's degree in Automotive Engineering in 2004 and the Ph.D. degree in Information and System Engineering in 2009 from Politecnico di Torino, Italy. In 2005 he worked for Fiat Research Centre, Italy, in the field of active vehicle systems. In 2007 he spent a three-months visiting period in the Optimization for Engineering Center (OPTEC) of the Katholieke Universiteit Leuven, hosted by prof. Moritz Diehl. Lorenzo Fagiano currently holds a post-doctoral position in the Complex Systems Modeling and Control group of the Department of Control and Computer Engineering of Politecnico di Torino, Italy. His main research interests include high-altitude wind energy generation using controlled tethered wings, constrained robust and nonlinear control, set membership theory for control purposes and automotive control systems. Lorenzo Fagiano is co-author of about 40 papers published in international journals, conference proceedings and book chapters. He is recipient of the ENI award "Debut in Research" prize 2010 and of the Maffezzoni prize 2009. He is member of IEEE, IEEE Control Systems Society, SIAM and SIAM activity groups on Optimization and on Control and Systems Theory.