Papers

13

Total Citations

443

H-Index

9

About

Boris Sofman is a leading figure in field robotics, whose work has fundamentally advanced how autonomous vehicles navigate unstructured, outdoor environments. His research centers on self-supervised learning, terrain perception, and real-time anomaly detection, all aimed at making robots safer and more capable in the wild. In his highly cited 2006 paper (113 citations), Sofman pioneered a self-supervised online learning framework that allows robots to improve their navigation by leveraging overhead imagery—a feature that traditionally generalized poorly—to predict terrain traversability. This work laid the foundation for his broader contributions to long-range navigation, where he demonstrated how aerial data could be processed to support ground vehicles over vast distances. Sofman also made critical advances in robot safeguarding, developing "anytime" novelty and change detection systems that enable robots to identify hazardous, unfamiliar situations before failure occurs. His 2010 paper on learning for autonomous navigation (94 citations) synthesizes these themes, showcasing how machine learning can bridge the gap between prior knowledge and onboard perception. With over 400 total citations, Sofman’s research remains essential reading for anyone building robots that must operate reliably in complex, unpredictable terrain.

Research Focus

Key Achievements

9
H-Index
13
Papers
443
Total Citations
34
Avg Citations/Paper
🏆 Most Cited Paper
Improving robot navigation through self‐supervised online learning
113 citations · 2006
📈 Most Prolific Year: 2010 (4 Papers)
🤝 Key Collaborators: 14
🏛 Institutions: Carnegie Mellon University, Sandia National Laboratories California

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