Generalized H-infinity Vehicle Control Utilizing Dimensional Analysis

The selection of a nominal plant model is a central design choice for a robust controller design. In the case of vehicle dynamic studies, the nominal plant model for a vehicle is traditionally chosen using knowledge of a particular experimental vehicle under study. With such a design focus on one particular plant, it may be questionable whether the resulting controller synthesis technique provides experimental conclusions that are generalizable to several other vehicle types. This work develops an alternative, dimensionless representation of vehicle dynamics that is more suitable for a generalized vehicle analysis. Within such a nondimensional framework, the average of vehicle parameters becomes well defined, and perturbations about the average are easily developed that reasonably encompass all production vehicles. These uncertainty bounds are then used to generate a robust controller suitable for nearly all passenger vehicles. For the purposes of demonstration, the focus of this work is a lateral-positioning control task. The resulting controller is demonstrated on a scaled experimental vehicle.