Data-driven Dynamic Relatively Optimal Control
Date:
I presented the paper “Data-driven Dynamic Relatively Optimal Control” on AUTOMATICA.it 2022
Content:
We show how the recent works on data driven open-loop minimum-energy control for linear systems can be exploited to obtain a stabilizing state feedback compensator (actually deadbeat) (i) that does not require feedforward or state initialization, (ii) which relies only on collected experimental data (without the need for a mathematical formulation of the model), and (iii) that is optimal for the nominal initial condition. The result is a data-driven dynamic relatively optimal controller (ROC). As an example, we apply the proposed approach in a point-to-point cart and pole control task. The results show the effectiveness of the proposed data-driven dynamic ROC.