Data-based variable universe adaptive fuzzy controller with self-tuning parameters

Abstract

Data-based variable universe adaptive fuzzy controllers (VUAFCs) with self-tuning parameters are developed for complex systems with unknown universes in this paper. The main feature of the proposed VUAFC is the new defined contraction–expansion (C-E) factor on an infinite universe, which is more flexible and practical in real applications. Moreover, the data-based methods to tune the parameters including the peak points of output fuzzy subsets and the C-E factor parameters are proposed. The peak points of the output fuzzy subsets are mined by an improved Wang–Mendel method based on conflicting rules. The parameters of the VUAFC are optimized by solving an offline optimization problem using the chaotic particle swarm optimization (CPSO) algorithm. The simulation results on the strip temperature control of the radiant-tube indirect-fired furnace of annealing furnace show that our proposed method has strong practicability and good control performance.