Support vector regression based 3-D fuzzy logic controller design for spatially distributed systems

This paper presents a data-driven based 3-D FLC design methodology using support vector regression (SVR) learning. The key technique is to relate spatial fuzzy basis functions of a 3-D FLC to kernel functions of a SVR and construct an equivalence relationship of a 3-D FLC and a SVR. Therefore, a 3-D FLC can be established using the learned results of a SVR. Utilizing the concept of reference function, 3-D membership functions can be generated through location transformation. Thus, we can have more choice for 3-D membership functions. The proposed method was applied to a catalytic packed-bed reactor and simulation results have verified its effectiveness.