A Control Structure for Adaptive Interactive Systems

The adaptive interactive systems are still in the beginning of their practical use.There are many scientific prototypes but only few real-life applications usingadaptive technology. A lot of Internet applications use profiling techniques butthey are still parameter scheduling systems, which are not really adaptive. Theadaptive interactive systems have large-scaled very complex structures withhigh number of parameters. Great number of these parameters can be biased andvague.There is a lack of a framework that supports the structured development ofadaptive systems. The control theory is appropriate to model the structure andparameters of such systems. It allows the study of the system stability and thedetermination of parameter values for which the interactive systems are stableand optimal. The objective of this paper is to present an adaptive controlstructure that is implemented in a MATLAB environment (Part-Enander andSjoberg 1999). Its neuro-fuzzy adaptation feedback is very appropriate formodelling the uncertain system parameters.