Papers

15

Total Citations

186

H-Index

8

About

Michael Fulton is a robotics and computer vision researcher whose work sits at the intersection of marine robotics, underwater human-robot interaction, and environmental AI. He is perhaps best known for developing the **TrashCan** dataset series — a comprehensive, semantically segmented collection of underwater debris imagery that has become a foundational resource for marine litter detection research, accumulating over 57 citations since its 2020 release. His early contributions, including Trash-ICRA19, helped establish the data infrastructure necessary for training robust deep learning models in challenging underwater environments. Beyond environmental monitoring, Fulton has made significant strides in enabling meaningful collaboration between autonomous underwater vehicles (AUVs) and human divers. His work spans diver detection using deep neural networks, motion prediction for safer robot navigation, and novel communication frameworks — including gestural languages and the innovative SIREN audio system, which transforms a robot's outer hull into a speaker. His 2022 study on motion-based robot-to-human communication further extended these ideas to field robotics more broadly. With over 160 cumulative citations across his most prominent works, Fulton's research meaningfully advances both ocean conservation technology and the human-robot collaboration capabilities essential for real-world underwater operations.

Research Focus

Key Achievements

8
H-Index
15
Papers
186
Total Citations
12
Avg Citations/Paper
🏆 Most Cited Paper
TrashCan: A Semantically-Segmented Dataset towards Visual Detection of Marine Debris
57 citations · 2020
📈 Most Prolific Year: 2020 (4 Papers)
🤝 Key Collaborators: 14
🏛 Institutions: University of Minnesota, University of Minnesota System, Twin Cities Orthopedics

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 · 14 days ago