Previously (up until FS22) named "Control Systems II"
Abstract
The focus of the course is on the design of advanced controllers for cyber-physical systems, that is, systems in which the controller is an embedded computer that can sense and actuate a physical plant. Advanced computational control strategies like Model Predictive Control, Reinforcement Learning, and Data-Driven control will be covered.
Objective
The objective of the course is to prepare students to the design of advanced digital control systems: this includes comparing alternative control strategies, deciding what class of controllers to employ for a specific problem, tune the controller in order to meet the desired specifications, and produce a conceptual design of how the controller can be implemented and deployed. Simplifying assumptions on the underlying plant that were made in the course Control Systems are relaxed, and advanced computational control concepts and techniques are presented.
Content
The course will cover both the challenges of a digital control system and the many possibilities offered by powerful computation in control. Different aspects and challenges of embedded control of cyber-physical systems will be discussed. We will then review the limitations of classical control strategies like PID control and LQR control, and motivate the need for controllers that employ significant real-time computation. In particular, we will look into Model Predictive Control, Reinforcement Learning, Data-Driven control, and possibly other advanced computational control techniques.
Lecture notes
Lecture notes will be available on the Moodle page of the course.
Literature
References to the literature will be provided during the course. No textbook is necessary, but students are encouraged to read the suggested readings.
Prerequisites / Notice
Prerequisites: Control Systems or equivalent. A background in optimization is very helpful. Students that don’t have it will be provided with some additional reading material.
