Peter Krauthausen
Georgia Institute of Technology, Karlsruhe Institute of Technology
Papers
7
Total Citations
112
H-Index
5
About
Peter Krauthausen is a leading researcher in robotics and autonomous systems, with key contributions spanning distributed inference, human-robot cooperation, and intention recognition. His seminal work on multifrontal QR factorization for multirobot localization and mapping (2005, 30 citations) revolutionized SLAM by demonstrating how factor graphs and rooted clique trees enable parallelized computation, significantly improving efficiency in distributed inference. This foundational approach, further refined in his work on exploiting locality in SLAM through nested dissection (2006, 19 citations), has become a cornerstone for scalable mapping algorithms. Krauthausen is also widely recognized for his pioneering multi-level framework for intention, activity, and motion recognition in humanoid robots (2011, 27 and 22 citations), which integrates monocular vision with domain knowledge for real-time, online operation. His subsequent research on situation-specific intention recognition and model-predictive switching (2010, 5 and 4 citations) advanced human-robot cooperation by enabling efficient, uncertainty-aware intention estimation using dynamic Bayesian networks. Through these contributions, Krauthausen has profoundly impacted how robots perceive, reason, and collaborate with humans in complex, real-world environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4Exploiting Locality in SLAM by Nested Dissection19 citations · 2006
- 5Situation-Specific Intention Recognition for Human-Robot Cooperation5 citations · 2010
- 6
- 7A model-predictive switching approach to efficient intention recognition4 citations · 2010