A novel robust position estimator for self-sensing magnetic levitation systems based on least squares identification

In this work a novel method is introduced for the estimation of the position of a self-sensing magnetic levitation system, based on a least squares identification strategy. In the first step, a detailed mathematical model of the magnetic levitation system is derived and the properties of this system are analyzed for the case of a pulse-width modulated control. Based on this model, an estimation algorithm for the inductance of the magnetic levitation system is introduced. In classical position estimation schemes known form the literature large estimation errors are typically induced by a deviation of the electric resistance from its nominal value or by a fast motion of the levitated object. In this work it is shown that these errors can be exactly compensated by means of a suitable estimation strategy. Furthermore, it is outlined that the chosen structure of the estimation scheme allows for a very efficient implementation in real-time hardware. Afterwards, the design of a cascaded position controller for the magnetic levitation system is briefly summarized. Finally, the excellent quality and the high robustness of the proposed position estimator is demonstrated by means of simulation studies and measurement results on a test bench.