Integration of Bilinear Systems and Neural Networks for Designing Nonlinear Semi-Active Suspensions

This paper presents a new design method of semi-active suspensions based on the integration of neural networks and bilinear systems. It is known that semi-active suspensions with ideal linear components have a bilinear structure. However actual semi-active suspensions with nonlinear components have an structure which is not purelv bilinear. In order to improve the performance of semi-active suspensions, neural networks and bilinear systems are integrated and used Tor the identification and optimal control of nonlinear semi-active suspensions. The validity and applicability of the proposed method are analyzed and verified theoretically and experimentally using a semi-active suspension model equipped with piezoelectric actuators.