A new notion of robust $\mathit{F}$-stability is introduced, based on the recently introduced notion of $\mathcal{F}$-passivity. Necessary and sufficient conditions for the robust $\mathit{F}$-stabilization under dynamical structured uncertainties are given, in terms of matrix inequalities. If besides dynamical, also statical uncertainties are included, the given conditions are only sufficient. The conditions are not convex, but for two dual classes of plants, it is shown that the conditions are convex, expressed by linear matrix inequalities (LMIs). An algorithm for controller design is given, and is illustrated by academic examples. It is shown that the uncertainty region where the closed-loop system is stable, obtained by the algorithm, is maximal in respect to literature. It is shown that some of the uncertainties/channels can be completely rejected using the control. If the design algorithm with a transformation is applied to simultaneous robust stabilization and robust ${\mathscr{H}}_{\infty }$ performance, it improves the performance that can be obtained using the method based on inclusion of an extra uncertainty block in the purely robust stabilization scheme.

稳健的新概念$\mathit{F}$-基于最近引入的概念，引入了稳定性$\mathcal{F}$-被动。稳健的充要条件$\mathit{F}$-根据矩阵不等式给出了动态结构不确定性下的稳定性。如果除了动态的，还包括静态的不确定性，给定的条件就足够了。条件不是凸的，但对于两个对偶类别的植物，表明条件是凸的，由线性矩阵不等式 (LMI) 表示。给出了一种控制器设计算法，并通过学术实例加以说明。结果表明，该算法得到的闭环系统稳定的不确定区域在文献中是最大的。结果表明，使用控制可以完全拒绝一些不确定性/通道。如果将带有变换的设计算法应用于同时鲁棒稳定和鲁棒${\mathscr{H}}_{\infty }$性能，它提高了使用基于在纯鲁棒稳定方案中包含额外不确定性块的方法可以获得的性能。