Modeling and output feedback control of a floating ball inside a tube

This thesis presents modeling, simulation, and real-time output feedback control of a floating ball system that contains a noisy sensor feedback. It concentrates on the control of the vertical position of an object inside a tube pushed upward by an air flow. The procedures of this thesis are to obtain a dynamic model, introduce an effective filtering method, and implement the filtering method on the system in real-time.

The improvement of the floating object's convergence to a desired vertical position is demonstrated by comparing two control techniques with and without the filtering method. First, the results of the convergence of the system output are shown by using a Fuzzy Logic Controller. Then, the new convergence results are presented by using the Fuzzy Logic Controller with the Median Filter. Second, the convergence results of the system output are shown using the tracking controller. Then, again, the new convergence results are shown using the tracking controller with the Median Filter.

Removing the noise of the feedback in a floating ball system leads to less of the unwanted oscillatory response from the system, faster settling time of the floating object to a desired height, and disturbance rejection. This work may further improve an object's floating stability at a desired altitude of a floating object system containing a very noisy sensor feedback.