Improvement of adaptive neuro-fuzzy controller using by fuzzy clustering means algorithm for control of vehicle suspension system

Suspension system is an important part of vehicle whose main role is to separate the vehicle body from road induced vibrations. Design and control of a suspension system that can adapt to different road conditions with high flexibility is essential. In this study, data were collected from three types of road conditions with different roughness coefficients in various forward speeds for training a suspension model. Primarily, dynamic equations were derived for a linear full model suspension system. Then, with the use of fuzzy system simulation data, two adaptive neuro-fuzzy controllers namely Grid Partitioning and Fuzzy Clustering were trained. Finally, four methods were evaluated and the results showed that decrease in linear deflection and acceleration of vehicle body is higher in adaptive neuro-fuzzy controller by Subtractive Clustering compared to other systems.