Application of pattern recognition methods to control systems

Abstract

Pattern recognition methods, capable of advanced decision making without a precise description of the process model, have been used in many cases to implement advanced control systems. Such techniques have been applied to structural identification of nonlinear systems, direct control of systems operating in unfamiliar environments, and quality process control. Finally, pattern recognition methods may be used to implement the higher levels of "intelligent control systems" where advanced decision making is required. An account of these techniques and their application to controls is given in this paper. Possible new applications to socioeconomics, bioengineering, transportation and robotic systems, man-machine interactive systems, and other systems useful in advanced automation are suggested.