Fuzzy rules acquisition and parameters evolution based on fuzzy neural networks

Some methods are proposed for fuzzy rules acquisition and fuzzy system parameters tuning based on fuzzy neural networks. The feasibility of the proposed methods is tested with an experiment of automatic train operation simulation. This experiment is also used to compare the learning and control of fuzzy inference system with those of standard BP networks and basic fuzzy systems. A summary is made of the characteristics of the methods. The final result indicates that the methods of fuzzy rules generation and fuzzy system tuning are very effective