Koichi Ogawara

About

Koichi Ogawara is a pioneering researcher in robot learning and human-robot interaction, with a particular focus on the "Learning from Observation" (LFO) and "Programming by Demonstration" (PbD) paradigms. His work centers on enabling robots to acquire complex manipulation skills by observing human demonstrations, dramatically reducing the programming burden traditionally placed on robotics engineers. Among his most influential contributions is the development of sensor fusion techniques using Hidden Markov Models to recognize continuous human grasping sequences, a paper that has garnered 118 citations and remains a landmark in gesture recognition. Ogawara's research has also tackled the challenging problem of deformable object manipulation — particularly knot-tying tasks — breaking new ground in an area largely ignored by contemporaries, with related papers accumulating over 80 citations each. His framework for extracting essential interactions from multiple human demonstrations (96 citations) addressed critical ambiguities inherent in single-demonstration learning, advancing the robustness of robot task acquisition. His broader contributions span assembly task recognition, multi-fingered robotic painting systems, and attention-point analysis for task modeling. Collectively, Ogawara's body of work has shaped modern approaches to intuitive robot programming, making him a significant figure in intelligent robotics research.

Research Focus

Key Achievements

Top Papers

1. 1
   A sensor fusion approach for recognizing continuous human grasping sequences using hidden Markov models

   118 citations · 2005
2. 2
   Extraction of essential interactions through multiple observations of human demonstrations

   96 citations · 2003
3. 3
   Representation for knot-tying tasks

   82 citations · 2006
4. 4
   Knot planning from observation

   80 citations · 2004
5. 5
   Generation of a task model by integrating multiple observations of human demonstrations

   48 citations · 2003
6. 6
   Recognizing Assembly Tasks Through Human Demonstration

   46 citations · 2007
7. 7
   Painting robot with multi-fingered hands and stereo vision

   43 citations · 2008
8. 8
   Modeling manipulation interactions by hidden Markov models

   41 citations · 2003
9. 9
   Recognition of human task by attention point analysis

   25 citations · 2002
10. 10
    Acquiring hand-action models by attention point analysis

    22 citations · 2002

Key Collaborators

Contact & Links