A Game-Playing Fuzzy Logic Controller for Semi-Active Suspensions

The idea of controlling physical systems with artificially-intelligent control methods has great appeal, and has been investigated heavily in the past. At the very least, they offer the chance to control effectively systems that are highly nonlinear and/or time-varying. This paper describes attempts to fuse game-playing theory and binary-decision-tree pattern recognition methods to build a fuzzy logic parametric controller for a semi-active suspension system. The goal of the controller is to minimize (or defend against) the maximum possible transfer of road interaction forces to the vehicle. Simulation results show that the resulting system can provide the desired control actions to handle smoothly the irregularities in the road profile and the many nonlinearities inherent in the suspension system.