Moving least square online predictive model for two degrees of freedom suspension system using optimal adaptive fuzzy controller

The Moving Least Square (MLS) interpolation method is proposed for approximation of adaptive fuzzy controller parameters for two degrees of freedom suspension system and each one has two inputs, one output with twenty-five linguistic fuzzy IF-THEN rules. Fuzzy systems are designed by using five Gaussian membership functions for each input, product inference engine, singleton fuzzifier and center average defuzzifier. The constructed fuzzy systems is composed with adaptation rules. For this purpose, Lyapunove approach is implemented for stability of the adaptation rules. The Gravity Search Algorithm (GSA) is implemented for achieve the optimum controller parameters. The relative displacement between sprung mass and tire and the body acceleration are two objective functions used in the optimization algorithm. Since, choose the suitable controller coefficients are important and when the parameter of the system change, Optimum coefficients of the controller will also change. In order to solve this obstacle, the MLS predictive model is purposed that is interpolation method based on a radius of the neighborhood, a basis function and a weight function for points of interest. Finally online model is implemented on the two degrees of freedom suspension system and results compared with the offline optimal systems.