On the augmentation of Luenberger Observer-based state feedback design for better robustness and disturbance rejection

摘要

This paper addresses the fundamental problem of improving the Luenberger Observer (LO)-based feedback control so that it can deal with effectively a wide range of uncertainties, both internal and external. It is shown that the effect of such uncertainties can be estimated and cancelled in real time, all based on the yet untapped source of information: the estimation error in the existing LO. The result is quite significant in that, by augmenting the LO without changing its original structure or parameters, the LO-based design can be improved in terms of robustness and disturbance rejection. In other words, there is hidden information of uncertainties in the LO tracking error and it can be extracted, often with a mere integrator, to great benefits, as shown in both mathematical analysis and simulation studies.