Papers

49

Total Citations

1,249

H-Index

22

About

Ingmar Posmer is a leading researcher in robotics and machine perception, with key contributions in deep learning for tracking, scene understanding, and robust autonomous navigation. His work on end-to-end object tracking, notably "Deep Tracking" (180 citations), pioneered direct mapping from raw sensor data to object tracks using recurrent neural networks, eliminating the need for hand-crafted features. He also advanced object-centric generative models with the GENESIS framework (over 145 citations combined), enabling compositional scene inference critical for robotics and reinforcement learning. Posmer has addressed real-world challenges like appearance change in outdoor robotics through adversarial domain adaptation (66 citations) and developed introspective classification methods (42+ citations) that allow robots to know when they don't know—a vital capability for mission-critical decision-making. His research on efficient lidar-based localization (38 citations) and probabilistic prediction of perception performance (37 citations) further demonstrates his impact on reliable, large-scale autonomous systems. With a focus on bridging deep learning and practical robotics, Posmer's work has shaped how robots perceive, track, and reason in dynamic, partially observable environments.

Research Focus

Key Achievements

22
H-Index
49
Papers
1,249
Total Citations
25
Avg Citations/Paper
🏆 Most Cited Paper
Deep Tracking: Seeing Beyond Seeing Using Recurrent Neural Networks
180 citations · 2016
📈 Most Prolific Year: 2016 (8 Papers)
🤝 Key Collaborators: 105
🏛 Institutions: University of Oxford, Science Oxford, Oxford Research Group, Robotics Research (United States), Art Institute of Portland

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 13 days ago