Papers

5

Total Citations

57

H-Index

5

About

Jeroen van Baar is a robotics researcher whose work bridges the critical gap between simulation and real-world physical interaction. His primary research areas include deep reinforcement learning for robot control, sim-to-real transfer learning, and tactile perception. Van Baar’s major contributions lie in developing visual analytics tools for complex control policies—as demonstrated in his most-cited work, "DynamicsExplorer" (22 citations)—which enables researchers to better understand and debug LSTM-based control policies for dynamic tasks. He has also advanced sim-to-real transfer learning by introducing robustified controllers that allow policies trained in simulation to succeed in real robotic environments (15 citations). Beyond control, van Baar has made notable strides in tactile sensing, developing methods to synthesize volumetric meshes from vision-based tactile imprints (9 citations) and creating interactive tactile perception systems for novel object classification (5 citations). His work on visual-inertial odometry frameworks incorporating 3D points, lines, and planes further showcases his versatility in perception. With a growing citation record and publications spanning top venues, van Baar is establishing himself as a key contributor to making reinforcement learning practical for real-world robotic manipulation and perception tasks.

Research Focus

Key Achievements

5
H-Index
5
Papers
57
Total Citations
11
Avg Citations/Paper
🏆 Most Cited Paper
DynamicsExplorer: Visual Analytics for Robot Control Tasks involving Dynamics and LSTM-based Control Policies
22 citations · 2020
📈 Most Prolific Year: 2020 (2 Papers)
🤝 Key Collaborators: 15
🏛 Institutions: Mitsubishi Electric (Japan), Mitsubishi Electric (United States)

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 14 days ago