The Neuro-fuzzy Identification of MR Damper

It is extremely difficult to describe the direct and inverse model of the Magneto-rheological (MR) damper, because an MR damper has strong nonlinearity between inputs and output. The paper presents a novel way to model these two models by using the universal approximation of neuro-fuzzy system. Two different neuron-fuzzy systems are designed to identify the direct and inverse model on the basis of adaptive neuro-fuzzy inference system (ANFIS). The numerical simulation proves that such two neuro-fuzzy systems can precisely model the direct model and inverse model of the MR damper for the train data, and well approximate for the check data. This idea can be extended to other models of MR dampers and can be also used to control MR dampers.