ABSTRACT
In control, filtering, and prediction applications the available a priori model information is sometimes not sufficiently accurate for a fixed design to satisfy all of the design specifications. In these circumstances, the designer may
• Reduce the specified level of required performance. • Expend additional efforts to reduce the level of model uncertainty. • Design a system to adjust itself online to increase its performance level as the accumulated input
and output data are used to reduce the amount of model uncertainty during operation.
