Control problems arise in many applications. Wherever certain parameters of a dynamic system have to be kept constant or have to be changed to a desired value without a human but with the help of some measurements of the system's output, we say the parameters are automatically controlled. Well-known examples of control systems are the cruise control in a car, the autopilot in a plane, or an air-conditioning system.
In this lecture, we will learn how to design controllers if the system dynamics are non-linear, noisy, or not even available in a mathematical form, e.g., if the system is described by a large and costly numerical simulation of which only the binary file is available but not the source code itself. In this case, approaches from computational intelligence are well-suited to build and optimize a controller as they typically do not assume much about the system which is modeled/optimized (so-called "blackbox scenario").
The main topics of the lecture are:
Large parts of the lecture are based on examples and practical applications in which the presented ideas have to be implemented in MATLAB. A final written exam (open-book and partly on the computer) will determine whether the lecture material has been understood. Passing the exam is worth 2 ECTS points.
The lecture is held by Anne Auger and Dimo Brockhoff from January till March 2013 in a total of 8x3hrs.
| Anne Auger | Dimo Brockhoff | ||||||||||||||
|
|