On the design of knowledge based control systems in manufacturing
Abstract
The capability of a control system is largely dependant on its knowledge of the process. While analytical process models provide quantitative information about localized plant dynamics, they cannot represent the cause and effect behaviours typical of complex systems. Knowledge based, or 'expert' systems are capable of representing this information. There are however, considerations that must be made when applying such methodologies to control applications. This paper outlines the general requirements of knowledge based control systems, and with reference to these requirements, offers some design perspectives. Such perspectives are emphasized by a brief description of an application of a knowledge based controller for turning. The paper concludes with a discussion of further directions for research.