An on-line knowledge-based system for interpreting ultrasonic data

Describes a rule-based system (ARBS) which has been designed specifically for making on-line decisions on the basis of incoming signal data. The general facilities of the system are described, together with a prototype system for interpreting two dimensional (b-scan) images from manufactured components. ARBS uses a blackboard architecture to represent evolving knowledge. The system has transparent links with external procedures, which can be used to assist in interpretative and strategic decision-making. This facility is used extensively for pre-processing of the images and for numerically intensive calculations. A key feature of ARBS is its ability to offer an intimate mix of knowledge-based programming and conventional programming