Niklas Funk

Niklas Funk joined the Institute for Intelligent Autonomous Systems (IAS) at TU Darmstadt in September 2020 as a Ph.D. student.

Previously, Niklas received his bachelor‘s degree in Electrical Engineering and Information Technology, as well as his master’s degree in Robotics, Systems and Control from ETH Zurich. During his studies, he focused on the intersection between machine learning and control and completed several applied projects. He finished off his masters with the thesis – „Learning Event-triggered Control from Data through Joint Optimization“, which was conducted at the Max-Planck Institute for Intelligent Systems under the supervision of Dominik Baumann and Sebastian Trimpe, and has been awarded with the ETH medal.

Currently, Niklas is working in the broad field of autonomous robotic assembly. On the one hand, he is exploring algorithms enabling combinatorial reasoning and generalization. In particular, his recent research focuses on combining learned graph-based representations with powerful inductive biases and model-based search. On the other hand, he is also interested in developing fine, precise, and dexterous robotic assembly skills, eventually adding tactile feedback.

Research Interests

Reinforcement Learning, Control, Robotics, Graph-based Representations, Tactile Sensing

Key References

Reinforcement Learning and Combinatorial Optimization for Robotic Assembly

• Funk, N.; Chalvatzaki, G.; Belousov, B.; Peters, J. (2021). Learn2Assemble with Structured Representations and Search for Robotic Architectural Construction, Conference on Robot Learning (CoRL).   Download Article [PDF]   BibTeX Reference [BibTex]
• Funk, N.; Menzenbach, S.; Chalvatzaki, G.; Peters, J. (2022). Graph-based Reinforcement Learning meets Mixed Integer Programs: An application to 3D robot assembly discovery, IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).   Download Article [PDF]   BibTeX Reference [BibTex]
• Wibranek, B.; Liu, Y.; Funk, N.; Belousov, B.; Peters, J.; Tessmann, O. (2021). Reinforcement Learning for Sequential Assembly of SL-Blocks: Self-Interlocking Combinatorial Design Based on Machine Learning, Proceedings of the 39th eCAADe Conference.   Download Article [PDF]   BibTeX Reference [BibTex]

Dexterous Manipulation

• Funk, N.; Schaff, C.; Madan, R.; Yoneda, T.; Urain, J.; Watson, J.; Gordon, E.; Widmaier, F; Bauer, S.; Srinivasa, S.; Bhattacharjee, T.; Walter, M.; Peters, J. (2022). Benchmarking Structured Policies and Policy Optimization for Real-World Dexterous Object Manipulation, IEEE Robotics and Automation Letters, 7, 1, pp.478-485.   Download Article [PDF]   BibTeX Reference [BibTex]

Event-triggered Control

• Funk, N.; Baumann, D.; Berenz, V.; Trimpe, S. (2021). Learning event-triggered control from data through joint optimization, IFAC Journal of Systems and Control, 16, pp.100144.   Download Article [PDF]   BibTeX Reference [BibTex]

Supervised Theses and Projects

Year	Type	Together with	Student(s)	Topic
2022	Master Thesis	Georgia Chalvatzaki	Simon Kiefhaber	Leveraging Energy-Based Models for Multi-Object 6D Pose Estimation
2022	Bachelor Thesis	Pascal Klink	Leon Magnus	Real-time Object Tracking for Assembly
2022	Bachelor Thesis		Svenja Menzenbach	Leveraging Learned Graph-based Heuristics for efficiently solving the Combinatorics of Assembly
2021	Master Thesis	Boris Belousov	Frederik Wegner	Learning Vision-Based Tactile Representations for Robotic Architectural Assembly
2021	Integrated Project	Belousov, B.; Chalvatzaki, G.	Jan Emrich, Simon Kiefhaber	Probabilistic Object Tracking Using Depth Carmera
2021	Integrated Project	Belousov, B.; Chalvatzaki, G.	Leon Magnus, Svenja Menzenbach, Max Siebenborn	Object Tracking for Robotic Assembly
2021	Integrated Project	Belousov, B.; Chalvatzaki, G.; Wibranek, B.	Jan Schneider	Architectural Assembly: Simulation and Optimization