Robotics talents tackle tricky issues
A humanoid robot learning to walk, a robotic arm playing draughts with itself, and a robot performing measurements in the micrometre range: During the interim presentation of their projects in the bachelor's degree programmes Robotik and Robotics at the Technical University of Applied Sciences Würzburg-Schweinfurt (THWS), students proved that they had tackled numerous technical issues.
The event during which the students presented their interim results in the large lecture room on Campus Konrad Geiger was well attended. The development project is a key feature of the bachelor's degree programmes Robotik and Robotics. It takes place in the sixth and seventh semester. In this projects, students can apply everything they have learnt so far: programming software, installing hardware, and of course, working together in a team. The topics are suggested by professors and lab staff. They can also be commissioned by companies cooperating with THWS.
One of the student groups supervised by Professor Dr. Christian Ziegler decided for one of these cooperation projects: Commissioned by Schaeffler AG & Co. KG, the students are working on a challenging problem – automated handling of ball bearing components including measurements in the micrometre-range, i.e. accurate to one thousandth of a millimetre. This accuracy, regarding the roundness of the balls as well as the circumference of the ball bearing rings, is essential so that the ball bearing can run smoothly. This work step is currently done manually millions of times in the ball bearing city of Schweinfurt. Employees manually insert the rings into a measuring device and turn them to measure their circumference with an accuracy of one millionth of a metre. Until now, this monotonous work process could not be automated, as the sensitivity of the human hand could not yet be replaced.
"In addition to programming, the heart of the smart robot solution is a specially developed tool that enables the automated transfer of the rings to the measuring system," explains Professor Dr. Ziegler. "An integrated, rubber-coated drive wheel, in combination with the robot's sensor technology, ensures that the ring can be rotated with high precision in the measuring system without causing unwanted measurement errors." The students not only successfully applied the content from lectures and exercises, but also continuously questioned and optimised their concept, says Professor Dr. Ziegler. "The student group is currently working on a specific challenge from everyday production: What happens when oil used in the manufacturing process causes the rings to slip through the measuring system when turning? Also for this issue, the group has already developed a solution idea, on which they are still working."
NAO learns new walking movements
The group supervised by Professor Dr. Stefan Friedrich chose a completely different topic: How does a robot learn to walk? First, the students developed a dynamic simulation model for the NAO-robot, with which they can try out any pattern of movement. "However, one thing to note is that the robot always falls over at the beginning – teaching a robot to walk is not that easy", says Professor Dr. Friedrich. But this would not matter in a simulation as the robot could try to start walking thousands of times. Taking into account the innumerable simulation runs, artificial intelligence is used to optimise the process and develop a successful walking pattern. "Finally, the motion control from the simulation is to be transferred to the real robot – the group is still working on this," Professor Dr. Friedrich explains.
Combination of AI and traditional engineering knowledge
The team of lab engineer Rodolfo Verde Castillo accepted the challenge of teaching a robot how to play draughts: "The biggest challenge was for the robot to reliably recognise the game pieces, grab them, and put them down. For this, the group has developed an automated calibration system that recognises the height of the game board and afterwards takes the borders as an orientation," Verde Castillo explains. Due to this calibration, the robot can handle slight changes in position of the game board without human assistance. The robot must also be able to tell apart the two different colours of the game pieces.
"The students have found an elegant solution for turning a game piece into a queen," Verde Castillo praises his team. "They built a flip mechanism that allows a game piece to be turned over so that the queen side is facing up." To ensure that everything fits together flush, the game pieces and the game board have built-in magnets. Currently, two different algorithms are running against each other in the robot's "brain", so that when playing draughts, two AIs play against each other. The team is currently working on the next step, which will involve the robot competing against humans.
A total of 13 teams showed their progress during the interim presentations, highlighting the broad range of topics from droned to automated racing cars. The students have until the end of the winter semester to put the finishing touches to their projects.
"It is fascinating to see the students realise in the course of the project that they can actually apply their theoretical knowledge from their studies to tackle real-life practical problems," says Professor Dr. Dorit Borrmann, organiser of the presentation. "Along the way, students often encounter unexpected problems. However, the prospect of a working result spurs their ambition to find creative solutions."
More information on the bachelor’s degree programme Robotik and its English equivalent, the TWIN programme Robotics at THWS
About THWS
The Technical University of Applied Sciences Würzburg-Schweinfurt (THWS) is one of the largest universities of applied sciences in Bavaria and has driven teaching and applied research since it was founded in 1971. With about more than 9,100 students, a wide range of more than 60 degree programmes, and two doctoral centers, THWS covers a broad spectrum ranging from technology to economics and social sciences, language and design. THWS is not only rooted in the region, but also has a strong international focus, which is reflected in numerous collaborations and exchange programmes worldwide and not least in a versatile English-taught degree programmes.
Contact:
Technical University of Applied Sciences Würzburg-Schweinfurt
Center for Robotics (CERI)
Prof. Dr. Dorit Borrmann
Ignaz-Schön-Str. 11
97421 Schweinfurt
Press contact:
Technical University of Applied Sciences Würzburg-Schweinfurt
Angela Kreipl
Münzstr. 12
97070 Würzburg
+49 931 3511-8354
