Real systems are often nonlinear and cannot be controlled by classical methods. It is necessary to use model-based control. A wide range of algorithms and approaches that work with the system model in the design phase of control or even in real time can be included in this category. The research group deals with the following four interrelated areas: model structure design, model parameter estimation, state estimator design, and controller structure and parameter design. Within these areas, we have long been developing the application of local linear models, the use of gradient, evolutionary and other optimization methods, the deployment of parallel computing, GPU computing and also hybrid controller structure (combination of feedback and direct-coupled controller). Furthermore, we address the issues of applicability of these algorithms and methods on practical HW, including their adaptive variants. The theoretical approaches have been applied to the control of such systems as a BLDC drive of an aircraft fuel pump, a positionally controlled satellite antenna, and a strongly nonlinear actuator in aerospace and automotive applications.
Projects
National Competence Centre of Mechatronics and Smart Technologies for Mechanical Engineering
Ultra high-speed electric active brakes for testing electric vehicle powertrains
National Competence Centre for Aeronautics and Space
Significant publications
Contact person
doc. Ing. Robert Grepl, Ph.D.