Adaptive Interactive Systems
Alexander Nikov, Tzanko Georgiev Technical University of Sofia, PO Box 41, BG-1612Sofia, Bulgaria email: nikov@vmei.acad.bg
Hans-Günter Lindner humanIT Human Information Technologies GmbH, GMD, TechnoPark, Rathausalle 10, 53754 Sankt Augustin, Germany email: lindner@humanit.de
The adaptive interactive systems are still in the beginning of their practical use. There are many scientific prototypes but only few real-life applications using adaptive technology. A lot of Internet applications use profiling techniques but they are still parameter scheduling systems, which are not really adaptive. The adaptive interactive systems have large-scaled very complex structures with high number of parameters. Great number of these parameters can be biased and vague.
There is a lack of a framework that supports the structured development of adaptive systems. The control theory is appropriate to model the structure and parameters of such systems. It allows the study of the system stability and the determination of parameter values for which the interactive systems are stable and optimal. The objective of this paper is to present an adaptive control structure that is implemented in a MATLAB environment ( Pärt-Enander and Sjöberg 1999). Its neuro-fuzzy adaptation feedback is very appropriate for modelling the uncertain system parameters.
-351-