Model-Based Control of a Large-Scale Ball-on-Plate System With Experimental Validation

A ball-on-plate system is a widespread education oriented laboratory experiment for automation in mechatronics. The setup combines elements of mechanical, electrical and control engineering and is an adequate setup for learning the combination of theory and practice. This paper presents an example of a workshop result on automatic control in mechatronics. The aim of the workshop is to develop and compare model-based approaches for ball position control in a given large-scale ball-on-plate system. The result includes the derivation of a non-linear state space model of the system. The equations are linearized in the center of the horizontal plate as an operating point in order to apply cascade control, linear-quadratic optimal control and PI optimal state feedback control. The algorithms are implemented on a microcontroller and tested in the experimental setup. The results show a successful control development which achieves the control goal with good performance in terms of command response.