Combining real-time with knowledge processing techniques

Abstract

Real-time systems make high demands on the programming style such as timeliness and predictability. The architecture which is presented in this paper uses explicitly given knowledge to meet the requirements of next generation real-time systems. For this purpose a multi-stage architecture for real-time knowledge processing systems was developed. It is based on a simple knowledge model. This can be implemented very efficiently by using the ARON-technique (Alternatives Regularly Organized and Numbered), which is easy to use even by non-programmers. Starting with this elementary system, more complex applications can be handled by decomposing the solution of a problem into interacting subsystems, each working with the ARON-technique. Applicability is demonstrated by two examples of knowledge based process control, e.g. real-time control of an A.C. motor.