OPTEC - K.U.Leuven Center of Excellence: Optimization in Engineering

 Nonlinear Model Predictive Control (NMPC) Tutorial Workshop

 Lecturers: Rolf Findeisen, Niels Haverbeke, Moritz Diehl

Course website

Aim of this 5 hour intensive workshop is to provide the participants with
 knowledge about the basic concepts of nonlinear model predictive control
 (NMPC). It is divided into five lectures, ranging from an introduction, over
 stability theory, numerical solution, state estimation, and current research
 topics. The workshop in particular profits from the presence of Rolf
 Findeisen (Stuttgart University), an internationally recognized NMPC expert,
 who visits OPTEC from Jan 16-18 and will give three of the five lectures. The
 workshop can be quite interactive and participants are expected to ask many
 nasty questions in the long coffee breaks ;-)


 WORKSHOP PROGRAM AND ABSTRACTS:

 13:30: Rolf Findeisen: Introduction and overview to NMPC (Slides)

 The main focus in this lecture is laid on an introduction and historical
 perspective of (nonlinear) predictive control. Specifically we outline the
 basic principle of predictive control, reasons for the huge success of linear
 model predictive control and the key advantages, disadvantages and challenges
 inherent in NMPC.

 14:15: Rolf Findeisen: Basic theory and stability of NMPC

 Nonlinear model predictive control is based on the repeated solution of a
 (finite) horizon open-loop optimal control problem subject to system dynamics
 and input and state constraints. However, as is well known by now, optimality
 does not automatically imply stability in the case of finite prediction
 horizons. Different approaches to achieve closed-loop stability using finite
 horizon lengths exist. The main purpose of this lecture is to review the
 underlying main ideas and theoretical foundations for these approaches.

 15:00: first coffee break

 15:15: Moritz Diehl: Efficient Numerical Optimization for NMPC (Slides)

 A necessary prerequisite for NMPC is the fast and reliable solution of
 nonlinear optimal control problems in real-time. In the first part we give an
 overview of general optimal control methods - dynamic programming, indirect
 and direct approaches, with a focus on the last class. In the second part, we
 outline crucial ideas behind real-time iteration algorithms, and present some
 examples to illustrate their performance.

 16:15: Niels Haverbeke: Introduction to Moving Horizon Estimation for
 Nonlinear Systems (Slides)

 The goal of state estimation is to reconstruct the state of a system from
 process measurements and a model. In practice state estimators must address
 many different challenges including nonlinear dynamics and hard constraints
 on states or disturbances. The framework of moving horizon estimation (MHE)
 is well suited to address these challenges and in many cases MHE gives
 accuracy that is superior to the Extended Kalman filter which is widely used
 in a large range of application areas. This talk aims at giving a general
 introduction to MHE. We will highlight the key ingredients (stability,
 smoothing, …) and show its practical use in a feedback control scheme (e.g.
 in cascade with MPC).


 17:00: second coffee break

 17:30: Rolf Findeisen: Challenges and opportunities in predictive sampled data open-loop feedback control

 In the first part we will discuss the issues of output feedback stabilization
 using NMPC, since often the full state information as required for the
 prediction is not available. For this purpuse we outline different
 possibilities to achieve stabilizing output feedback control. In the second
 part we focus on the question of control over networks subject to delays and
 possible package losses. We will outline how these effects can be
 counteracted and compensated, while guaranteeing stability.

 18:30: End of the workshop