Controller design using white box and black box methods for a magnetically suspended robotic system

Formulating vertical position control algorithms for magnetically levitated motion is a difficult task. One of the major problems inherent in identifying the dynamic model is that it is unstable for open-loop control, making it difficult to properly estimate the system dynamics. One proposed method for handling this problem is to first develop a simplified white box model based on magnetic levitation with one electromagnet. A simple controller can then be designed and experimental measurements can be obtained. Black box methods can then be used on the experimental data to estimate the closed loop transfer function, from which it is possible to extract the actual dynamic model of the system. This second model can then be used in order to derive a better control system or used as a component of a state-space model for generalized three-dimensional motion.