Required in stage
First term
Second term
Both terms
Not organised
This year
Next year
Alternating years
External
Prerequisites
Taught by
Language of instruction
Duration
Identical coursesModel Predictive Control (B-KUL-H0E76A)
Cannot be taken as part of an examination contractAims
This course aims at presenting an overview of real-time optimization-based control of dynamical systems, also known as model predictive control (MPC). It presents system-theoretic properties of MPC, such as stability, invariance, offset-free control, regulation and tracking, as well as numerical algorithms for solving the resulting optimal control problems. The focus is on both linear and nonlinear, continuous-time and discrete-time systems in state-space form. A number of case studies is presented, ranging from attitude and navigation control of quadcopters, collision avoidance for autonomous vehicles and hybrid vehicle control to multiperiod portfolio optimization, power dispatch in smart grids.
Finally, the student will gain both a deep theoretical understanding of the main principles as well as practical experience with MPC through an assignment consisting of a series of theoretical exercises and an MPC design project applied to autonomous racing.
Previous knowledge
optimization, numerical linear algebra, basic systems & control theory
Is included in these courses of study
- Master in de ingenieurswetenschappen: wiskundige ingenieurstechnieken (Leuven) 120 ects.
- Courses for Exchange Students Faculty of Engineering Science (Leuven)
- Master of Mathematical Engineering (Leuven) 120 ects.
- Master of Electrical Engineering (Leuven) (Information Systems and Signal Processing) 120 ects.
Activities
2 ects. Model Predictive Control: Lecture (B-KUL-H0E76a)
Content
- Introduction to Optimal control modeling for control; state-space models; discrete-time optimal control; linear & nonlinear optimal control; dynamic programming; direct methods for optimal control.
- Model predictive control receding horizon principle; Lyapunov stability; constraint satisfaction & invariance; tracking and offset free MPC; robust & stochastic MPC; modeling hybrid systems and logic.
- State estimation (extended) Kalman filtering; moving horizon estimation; output feedback MPC.
- Numerical Optimal control active set & interior point methods; sequential quadratic programming; augmented Lagrangian methods; proximal algorithms; mixed-integer optimization.
2 ects. Model Predictive Control: Exercises and Laboratory Sessions (B-KUL-H0E77a)
Content
The sessions consist of exercises on the topics from the lectures. An assignment of a simulation based project providing practical experience with MPC using the tools from the exercise sessions is given during the first half of the semester. This assignment will be graded.
Evaluation
Evaluation: Model Predictive Control (B-KUL-H2E76a)
Explanation
The grading consists of two parts: a written exam (theoretical) and a grade for the assignment based on a written report.