This experimental unit forms part of a series of teaching systems developed in collaboration with the Department of Automation and Information Technology at the Harz University of Applied Studies and Research.
A ball-plate model acts as a weakly-coupled mechanical multivariable system. A fuzzy control is used to move the ball to a specific position quickly and with as little movement of the plate as possible, even when the position of the ball is modified by external influences.
The position of the ball is measured without feedback using a touch panel and the crisp signals sent to the fuzzy controller, where the signals are transformed into fuzzy input values and inferenced before being transformed back into a crisp output value. Two servo motors act as actuators during this process. The inclination of the plate is modified by the movements of the respective motors; these movements are transferred to the plate by the drive rod.
The learning contents of RT 123 are based on the fundamentals of RT 121. The RT 123 is a multivariable system with two separate fuzzy controllers, which can also be coupled. Optimisation of the system by fine tuning the parameters will be looked at in a later exercise.
The control algorithms are initially written in the user-friendly development software FSH-Shell, simulated and then compiled to generate microcontroller code. The control strategy can be optimised at a later date.
A joystick can be used to control the system manually. This allows the degree of difficulty of the control process to be estimated very accurately.