The topic of this thesis work is the control of a complex, nonlinear and underactuated mechatronic mechanism, namely a “ball and beam system” manufactured by Amira so that the control involves different feedback law strategies. Nowadays, the role of underactuated systems (the feature of which is that they have less actuators than degrees of freedom) is increasing. Examples of such systems include helicopters, various land vehicles, mobile robots, weight handling equipment etc.
The most important parts of this device are a beam with variable angle and a ball rolling freely on it. The only input variable of the system is the torque applied to the beam, while there are two degrees of freedom, namely the angle of the beam and the position of the ball. Both are controllable, but the position is considered as the controlled variable. This system is located in the Intelligent robots laboratory of the department, so I was able to work on a real system besides the theoretical investigations during the realization of my thesis work.
The sensors and the actuator were already installed, my job was to control the position of the ball. Two different kinds of control strategies are implemented on the system: a cascade structure consisting of two PID controllers, and another cascade solution, containing a modern fuzzy controller in addition to the inner loop PID. In the case of every control loops, the thesis work introduces the identification of the model, the design of the controller, its implementation on the real system and the study of the performance and robustness of the control loop. A basic adaptivity feature is implemented, allowing our structure to control two different balls.
As for the software environment, I have been using Matlab/Simulink (Matlab R12 for the control itself and a newer version for additional tasks) and dSpace ControlDesk during my work.