A Control-Theoretic View of Intelligent Control
Graduate School of Science and Technology, Kumamoto University, 2-39-1 Kurokami, Kumamoto 860-8555, Japan
The term “intelligent” heralds new developments in many established fields with traditional assumptions in research and development. In the case of intelligent control, an attempt was first made to utilize methodologies such as model-based control theory, artificial intelligence and operations research, and recently to draw inspiration from nature, biology, and artificial life. Common methodology in both eras is promoted that is more accepting of heuristics and approximations than is the case with most research in control engineering. However, the methodology is less insistent on theoretical rigor and completeness. There is an inherent difficulty in succinct characterization of intelligent control. Many people have a common intellectual understanding of intelligent control, but some significant questions about common understanding have been yet to be unsolved such what is as the definition of intelligent control, and why intelligent control systems are generally structured in a hierarchical way. These questions should be solved before entering the second stage of an intelligent control framework. In this paper, we present the controltheoretic view of intelligent control technology. First, the need of a new approach is introduced based on the concept of control, and secondly typical intelligent control techniques are provided with example of inverted pendulum. Finally the definition and fundamental architecture of intelligent control are discussed and an example from a biped locomotion robot helps clarify the effectiveness of hierarchical architectures.
Copyright© 2000 by Fuji Technology Press Ltd. and Japan Society of Mechanical Engineers. All right reserved.