Hierarchical ANFIS Identification of Magneto-Rheological Dampers

Abstract

Magneto-rheological (MR) dampers, recently, have been widely utilized in many different areas of engineering for their high properties. There are two different kinds of problems for MR dampers, the direct model and the inverse one. It is difficult to express of the direct model of the MR damper for its high nonlinearity and hysteretic characteristics. It is much more difficult to get the inverse model of MR damper, which means the determination of the input voltage so as to gain the desired restoring force decided by the control law. When identifying the direct and the inverse model of MR damper with Adaptive Neuro-Fuzzy Inference System (ANFIS), there exists curse of dimensionality of fuzzy system. Therefore, it will take much more time, and even the inverse model may not be identifiable. The paper presents a hierarchical ANFIS to deal with the curse of dimensionality of the fuzzy identification of MR damper and to identify the direct and the inverse model of MR damper. The numerical simulation proves that the proposed hierarchical ANFIS can model the direct and the inverse model of MR damper much more quickly than ANFIS without more changing of identification precision. Such hierarchical ANFIS shows the higher priority for the complicated system, and can also be used in system identification and system control for the complicated system.