robust state-feedback control of uncertain lpv systems using integral quadratic constraints

A linear matrix inequality (lmi)-based algorithm is developed to design a robust state-feedback controller using integral quadratic constraints (iqcs) for an uncertain linear parameter varying system (lpvs). the uncertain lpvs is described by an interconnection of a nominal lpvs which is solely dependent on the measurable parameters and a block-structured uncertainty. the iqc approach is implemented to model the input/output behavior of the uncertainties. in general, the robust synthesis method and the iqc stability analysis for the uncertain lpvs lead to a non-convex problem and are solved by the iterative algorithms. however, in the proposed method, the problem is converted into a convex problem. therefore, the lpv synthesis for the nominal lpvs and the iqc analysis for handling uncertainties are performed simultaneously. consequently, without any constraints on nominal system matrices, the proposed method might achieve a better performance and less computational burden. furthermore, the object is to minimize the l 2 -gain, h ∞  control, when the closed-loop asymptotical stability is also guaranteed. the performance and effectiveness of the proposed method are demonstrated based on two examples

robust state-feedback control of uncertain lpv systems using integral quadratic constraints

a linear matrix inequality (lmi)-based algorithm is developed to design a robust state-feedback controller using integral quadratic constraints (iqcs) for an uncertain linear parameter varying system (lpvs). the uncertain lpvs is described by an interconnection of a nominal lpvs which is solely dependent on the measurable parameters and a block-structured uncertainty. the iqc approach is implemented to model the input/output behavior of the uncertainties. in general, the robust synthesis method and the iqc stability analysis for the uncertain lpvs lead to a non-convex problem and are solved by the iterative algorithms. however, in the proposed method, the problem is converted into a convex problem. therefore, the lpv synthesis for the nominal lpvs and the iqc analysis for handling uncertainties are performed simultaneously. consequently, without any constraints on nominal system matrices, the proposed method might achieve a better performance and less computational burden. furthermore, the object is to minimize the l 2 -gain, h ∞  control, when the closed-loop asymptotical stability is also guaranteed. the performance and effectiveness of the proposed method are demonstrated based on two examples