Discrete-time observer-based gain-scheduled output feedback controller design with simultaneous optimization of scaling matrices

Abstract

It is well known that Gain-Scheduled (GS) controllers are very effective for controlling plant systems whose characteristics vary with measurable parameters. Thus, there have been many reports on GS output-feedback controller design problem; however, in controller design for practical systems, multiple uncertainty blocks are often introduced to represent several design requirements. Then, the design problem of GS output feedback controllers and scaling matrices for the multiple uncertainty blocks become a non-convex problem. Furthermore, there are very few reports on the design method for observer-based GS output feedback controllers which have simpler structure than full-order H ∞ -type GS output feedback controllers. On this issue, we propose a design method for observer-based GS output feedback controllers with the simultaneous optimization of the scaling matrices in terms of dilated Linear Matrix Inequality (LMI). A numerical example demonstrates the effectiveness of the simultaneous optimization of GS controller gains and scaling matrices.