1 Description

This course offers the opportunity to work on research-related topics in the areas of

• Extended Reality
• Robotics
• State estimation, and
• Measurement and control systems.

The specific tasks in each semester are based on current research work at ISAS. The tasks are worked on by small groups of up to three students, by designing, implementing, and testing suitable hardware and/or software components. Each group is supported by a designated supervisor regularly giving personal feedback and advice. In addition, access to the necessary hardware and equipment at ISAS is provided. This includes, among other things:

• GPU-server
• Telepresence lab with HTC Vive, Microsoft Hololens and room-sized haptic interface (HapticGiant)
• 3D printers
• Mechanical and electronic workshops

The main goal of the course is the implementation of theoretical methods in real-world applications. An initial familiarization and brainstorming phase is followed by an extensive practical phase in which, in addition to the implementation work, project and time management skills are trained. In parallel, results are documented in the form of presentations and a final report.

This course is worth 8 ECTS. Students may optionally earn an additional seminar certificate in conjunction with this practical course. Information on this can be found in the description of “Seminar for Practical Course: Machine Learning and Intelligent Systems”.

1.1 Example Topics

The topics offered in a semester will be published as the registration opens. These are some examples from past semesters:

Estimation of the End-Effector Forces of a Hydraulic Excavator: For the remote control of an excavator, it is helpful to know the forces acting on the bucket. However, there are no suitable sensors for this purpose. In the course, a method was developed to estimate the end-effector forces based on hydraulic pressure data and joint angle sensors, and it was tested on a real 24-metric-ton excavator.
Development of a Classifier for Predictive Fault Detection: In industrial applications, gearboxes and bearings are subjected to high loads, and their failure results in costly downtime. As part of the course, a data-driven classifier was developed to predict failures. The necessary data was recorded using a test platform provided for this purpose.
Development of a Controller for the Feedback Loop of a Bulk Material Sorter: Bulk material sorters are an important component in many industrial processes and are inherently designed to achieve the best possible separation of good and bad particles. To be able to adjust the purity levels of such a sorter, a suitable multi-variable control system was designed and evaluated through simulation during the practical course.

1.2 Related Specializations

This course is suitable for the following specialization subjects:

• Theoretical Foundations
• Robotics and Automation
• Anthropomatics and Cognitive Systems

2 Requirements

In addition to having a strong interest in at least one of the projects offered, we require the following:

• Enrollment in a master’s program in Computer Science, Wirtschaftsinformatik, or Mechatronics (Other degree programs are not automatically excluded but may require special approval from the respective examination office.)
• A strong ability to work both in a team and independently
• Excellent German or English language skills
• Preferably programming experience in a high-level language (Python, Julia, C++ or similar)