Data-Based Control of Feedback Linearizable Systems

We present an extension of Willems' Fundamental Lemma to the class multi-input multi-output discrete-time feedback linearizable nonlinear systems, thus providing a data-based representation their input-output trajectories. Two sources uncertainty are considered. First, unknown linearizing input is inexactly approximated by set basis functions. Second, measured output data contaminated additive noise. Further, we propose approach approximate solution simulation and matching problems, show that difference from true bounded. Finally, results illustrated on example fully-actuated double inverted pendulum.