Weight optimisation in \(\mathcal H_{\infty}\) loop-shaping (Q2486098)

The appropriate choice of frequency weighting functions in classical \(H_{\infty}\) control is generally difficult, and very often the designer has to proceed by trial and error to match the expected closed-loop system specifications. The author proposes a more systematic technique for choosing weighting functions, using a particular formulation of the problem as a non-convex optimization problem. An algorithm is described that can help find loop-shaping weights and design \(H_{\infty}\) suboptimal controllers in a more systematic and less time-consuming way.