Intelligent feedback linearizing controller for an off-road electromechanical suspension system

Researchers at the University of Texas at Austin, Center for Electromechanics (CEM) have developed a new actively controlled suspension system for vehicles. UT-CEM's suspension replaces the standard spring and damper assembly with a spring and electromechanical actuator. The work contained in this document describes the evolution of an Intelligent Feedback Linearizing (IFL) Controller for this suspension that establishes an accurate, linear, baseline suspension response as a foundation for more advanced control techniques.; The first key component of the IFL controller is the intelligent parameter estimation algorithm. The intelligent estimation algorithm employs structured neural networks to learn key suspension parameters on line. The use of structured neural networks allows the system parameters to be learned in a manner that allows physical insight into the suspension's response at all points in the operating space. These estimates are used to compute force inputs that approximately cancel dynamic suspension forces. Experimental results show that the intelligent parameter estimation algorithm developed as part of this work can provide substantially better force cancellation than previous algorithms.; The second key component of the IFL controller is an output feedback control loop that causes the sprung mass portion of the suspension response to approximate that of a Type I Chebychev filter. Controllers are developed for both quarter and half car suspension models. In the case of the half car suspension, a controller is developed that decouples the vertical and pitch responses so that they can be designed separately. Furthermore, it is shown that it is possible to bound the passband cutoff frequency for the appropriate filter response based on allowable wheel travel and peak acceleration limits.