Abstract
	Information plays a key role in systems control and decisions. In general, information can be gathered through sensory measurements, man-machine interface and/or signal (data) communications. However, in many industrial processes, some information or knowledge can not be observed real-time through sensor measurement. As a result, this kind of internal information has to be constructed by a state estimator, or so-called "software based sensor", in terms of system observed inputs and outputs. To insure that the gathered information is of good quality and useful content, the signal processing, e.g. signal filtering and pattern quantization, may be used, if necessary [1].
	Many practical problems, arising in the manufacturing environment, related to complexity, ambiguity, incompleteness, and ill-formed structures are dealt with in Artificial intelligence (AI) efforts. Most manufacturing problems have these characteristics. The effective acquisition of domain-specific knowledge may be the key to the success of knowledge-based systems. This conclusion was confirmed also by solving of the steel manufacturing knowledge-based system. Problem solving was concern in the knowledge acquisition in the manufacturing context.