Bachelor/Master Theses - Ball in a Tube

Aerodynamic Levitation of a Ball in a Tube (ALB) – possible tasks for Bachelor- and Master-Theses

Keywords: Feedback-Stabilization, Multivariable Control, Nonlinear Control, Sensor fusion/Image processing, Machine Learning, Digital Twin

Experimental plants play a key role in the research and teaching of control theory. In the Aerodynamic Levitation of a Ball (ALB) system, the motion of a ping-pong ball in a tube is controlled by an airflow that is generated by a fan and a bypass valve. The plant contains the main properties of fast and marginally stable motion dynamics as well as multiple input and output variables.

The Control is realized by a MATLAB-Simulink-Model using a Raspberry Pi single-board computer to execute the compiled code. To exchange data with sensors and actuators, a Serial communication to an Arduino microcontroller is used. 

 


Ball In Tube

​Figure 1: Photograph of ALB 
plant (left), detail view on 
the valve (right)

The following table contains possible task fields regarding the ALB-plant.

Control-oriented Tasks:   Short description:
Feedback Stabilization BA/MA Stabilize the plant by the use of controller schemes with more than one degree of freedom
Multivariable Control BA/MA Use the fan and the valve simultaneously
Sensor Fusion MA Design of an observer; Modelling of noise
Machine-learning-oriented Tasks:   Short description:
Automated Testing BA Automated evaluation of controller parameters
Online Machine Learning MA Control the plant by (reinforcement) learning algorithms
Further possible Task fields:   Short description:
Digital Twin MA Create a digital twin of the plant
Image Processing BA/MA Determine position and speed of the ball by using a camera on Raspberry Pi
Hardware Extensions BA/MA Addition of further sensors; Interconnection of plants

Info + Contact

Required Lectures/Software Skills: ​ 

BA+MA: Regelungstechnik EiT, Digitale Regelung, Matlab/Simulink, ggf: Arduino (C/C++);

MA: Mehrgrößenregelung, Modeling and Simulation, SPE, ggf: Nonlinear Control, ASCT

Possible start:  expected from WS2020/21

M. Sc. Micha Obergfell, BB519,
micha.obergfell@uni-due.de

Request for a thesis

Graduation Cap 36 Request for a thesis