Initial Situation:

The handling and processing of cables and wires is currently characterized by manual activities due to their flexible material behavior. However, due to the ever-increasing complexity of cable systems in vehicle construction, there is also an acute need for action in industry to automate the process. As part of an automation project with a major German OEM, an automation solution is to be developed for cable harness production and assembly in the vehicle.

Scope of the thesis:

The objective of this thesis is to implement methods for real-time path adjustment based on in order to grasp a wire harness and avoid entanglement of the wires. Several approaches (e.g. Reinforcement Learning, Learning-by-Demonstration and Sim2Real with Nividia Isaac Sim) should be tested and compared.  The thesis is roughly divided as follows:

• Familiarization with ROS2 and programming of industrial robots
• Adaptation of the simulation environment and training of agents
• Improvement of algorithms to generate the parameters for position correction
• Transfer and validation on real systems

Benefits

• Hands-on experience in robotics development
• Exchange with other students at FAPS
• Insights into other areas of research
• Application-oriented work for career entry

What you should bring:

• Interest in AI-supported robotics and learning systems
• Experience in programming with Python/C++ and ROS2, as well as basic knowledge of machine learning
• Independent, structured, and scientifically sound approach to work
• German (C1) or English (C1)

Other notes:

• Start date can be immediate
• Remote work possible
• Work scope can be individually tailored according to interests
• Please apply with a current grade transcript and resume

Kategorien:

Kontakt: