Algebraic time-derivative estimation-theoretical foundations, extensions, and relationship with concepts from classical systems theory

Abstract
Many problems in control, diagnostics, and signal processing rely upon the estimation of time derivatives with respect to measured signals. Usually, these signals are known only from online sampled data and are subject to severe noise levels. A novel approach to time-derivative estimation, called the algebraic time-derivative estimation method, has been initiated by M. Fliess, H. Sira Ramirez, and co-workers. The intention of the talk is to give a comprehensive exposition of the theoretical basis in a classical operational calculus setting, show some extensions for noise attenuation, and to present some recent real-time applications of the method on a laboratory braking system. The seminar will conclude with the presentation of first results on the relationship of the algebraic approach with classical estimation concepts.

Biographical Information: Johan is visiting the EECS Control Group through October 2007. He has an office in 4123 EECS (6-3875): Email: [email protected]

Johann Reger obtained his diploma degree (Dipl.-Ing.) in mechanical engineering in 1999 and his doctorate (Dr.-Ing) in electrical engineering in 2004, both from the University of Erlangen-Nuremberg in Germany. Thereafter, he was postdoc research fellow with the mechatronics section at CINVESTAV-IPN in Mexico-City. From 01/2005 to 03/2007 he was director of the control laboratory with the institute of automation and control at the University of the German Armed Forces in Munich. Since 04/2007 he is visiting research scholar with the EECS at the University of Michigan. His primary research interests include nonlinear systems, tracking and online-trajectory planning, systems with internal dynamics, and passivity-based control of electrical drives and power converters.