Selflearning fuzzy controller with smooth transfer characteristic and guaranteed convergence

Abstract

The paper presents a user-friendly way to design a smooth nonlinear control surface. The method can be seen as a fuzzy control design tool, but can also be seen in the context of neural networks, Bspline basis functions, or simply as a tool for setting up an input-output map.The design process is composed of two steps. First, expert knowledge is used in a rule based manner to set up the main structure of the control SLUface. Second, an automatic learning algorithm is used to improve the control surface and to compensate for slowly time varying effects like the aging of the system.Applications are mentioned and the convergence of the learning algorithm is proved under real world conditions.

Modeling of a closed-loop controlled laser machining process with different control strategies

“…The learning algorithms for a fuzzy system are basically the same as for an input-output map, besides the characteristic curve is represented as a linear combination of membership functions. For the goal of a smooth characteristic the membership functions are defined as 3rd order B-splines [15]. To avoid 'logical inconsistencies' singletons for the output must be used.…”

Ein kennfeldorientiertes Konzept für Fuzzy-Regler

“…Die Regelschlußfolgerungen s k für die hier vorgestellten Fuzzy-Regler-Konzepte können automatisch erlernt werden, wie in [17] näher ausgeführt wird.…”