Semi-global robust nonlinear control: state-dependent scaling and computational aspects

This paper develops a computationally tractable approach to the analysis and synthesis problem of robust control of nonlinear systems with structured gain-bounded uncertainty. It is shown how state-dependence of scaling matrices can be utilized to reduce the conservatism of nonlinear systems analysis and synthesis. The robust performance problem with state-dependent scaling is characterized by linear matrix inequalities (LMIs). This paper clarifies to what degree these LMI conditions can actually guarantee robustness of the nonlinear system. Furthermore, the controller synthesis problem of expanding the region where required performance is achieved is formulated as a convex optimization