Papers
106
Total Citations
1,546
H-Index
19
About
Jun Takamatsu is a prominent robotics researcher whose work spans robot learning, autonomous navigation, human-robot interaction, and AI-driven robot control. He is perhaps best known for his foundational contributions to the "Learning from Observation" (LFO) framework, which enables robots to acquire manipulation skills by watching human demonstrations rather than requiring explicit programming. His early work on extracting essential interactions from multiple demonstrations (96 citations) and extending LFO to deformable object manipulation — including knot-tying tasks (82 citations) — established him as a leading voice in imitation learning and task representation. Takamatsu has also made significant strides in autonomous mobile robotics, with his 2019 ROS-based navigation system combining 2D LiDAR and RGB-D sensing becoming widely cited (115 citations) in the field. More recently, he has embraced large language and vision models, demonstrating how ChatGPT (92 citations) and GPT-4V (60 citations) can translate natural language and visual demonstrations into executable robot actions — work that places him at the cutting edge of foundation-model-driven robotics. His survey on robotic waste sorting (57 citations) further reflects a commitment to real-world societal impact, making his research portfolio both technically rigorous and practically meaningful.
Research Focus
Key Achievements
Top Papers
- 1ROS based Autonomous Mobile Robot Navigation using 2D LiDAR and RGB-D Camera115 citations · 2019
- 2
- 3
- 4Representation for knot-tying tasks82 citations · 2006
- 5Knot planning from observation80 citations · 2004
- 6
- 7Challenges for Future Robotic Sorters of Mixed Industrial Waste: A Survey57 citations · 2022
- 8
- 9Recognizing Assembly Tasks Through Human Demonstration46 citations · 2007
- 10A Gesture-Centric Android System for Multi-Party Human-Robot Interaction43 citations · 2013