Parametric Uncertainty Reduction in Robust Multivariable Control

This dissertation presents a method for reducing the number of parametric uncertainties used in the design of a robust H-infinity controller. The resulting controller is shown to meet robust stability and performance requirements in the presence of all the modeled uncertainties. The approach used involves grouping parametric variations affecting the same open loop eigenvalue, then scaling one or more of these variations to accommodate the eigenvalue change caused by all the parametric uncertainties. This method is effective in those cases where a large number of parametric uncertainties cause current computer aided design software to fail to find a robust controller for the plant.