Sampled-Data Controller Synthesis using Dissipative Linear Periodic Jump-Flow Systems with Design Applications

Abstract

In this paper, we will propose linear-matrix-inequality-based techniques for the design of sampled-data controllers that render the closed-loop system dissipative with respect to \textcolor{blue}{quadratic supply functions}, which includes passivity and an upper-bound on the system's $\mathcal{H}_\infty$-norm as a special case. To arrive at these results, we model the sampled-data control system as a linear periodic jump-flow system, study dissipativity in terms of differential linear matrix inequalities (DLMIs) and then convert these DLMIs into a single linear matrix inequality. We will present three applications of these synthesis techniques: 1) passivity-based controller synthesis, as found in teleoperations, 2) input-output-response matching of a continuous-time filter with a discrete-time filter (by minimizing the $\mathcal{H}_{\infty}$-norm of a generalized plant) and 3) a sampled-data controller redesign problem, where the objective is to find the best sampled-data controller, in the $\mathcal{H}_{\infty}$-norm sense, for a given continuous-time controller. We will show that synthesising sampled-data controllers leads to better closed-loop system behaviour than using a Tustin discretization of a continuous-time controller.