Lutz Frommberger
University of Bremen, FZI Research Center for Information Technology
Papers
14
Total Citations
127
H-Index
7
About
Lutz Frommberger is a leading researcher in the intersection of robotics, spatial cognition, and machine learning, with a particular focus on reinforcement learning for robot navigation. His work centers on developing qualitative spatial representations that allow robots to understand and navigate their environments more effectively, especially in unstructured or noisy terrains. Frommberger's major contributions include creating abstraction mechanisms that enable knowledge transfer across different navigation tasks, allowing robots to generalize learned behaviors rather than starting from scratch in each new environment. His 2008 paper "Learning to Behave in Space" (19 citations) introduced a landmark-based circular order representation that revolutionized how robots perceive space during reinforcement learning. His 2010 work on structural knowledge transfer (17 citations) formally characterized three facets of abstraction—aspectualization, coarsening, and conceptual classification—providing a foundational taxonomy for the field. Frommberger also contributed to real-time 3D mapping for walking robots like LAURON III, demonstrating practical applications in unstructured terrain. His research has accumulated over 100 citations, establishing him as a key figure in making autonomous robots more adaptive and intelligent through spatial abstraction and transfer learning.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Machine learning for interactive systems and robots16 citations · 2013
- 4
- 5
- 6Temporal logic for process specification and recognition9 citations · 2012
- 7
- 8
- 9
- 10