Hiroshi Tsujino
Honda (Japan), University of Southern California, Honda (United States)
Papers
55
Total Citations
1,401
H-Index
22
About
Hiroshi Tsujino is a pioneering researcher in robot audition, dedicated to giving robots the ability to hear and understand speech in noisy, dynamic, real-world environments. His central contribution is the creation of HARK (Honda Research Institute Japan Audition for Robots with Kyoto University), an open-source software system that enables robots to localize, separate, and recognize multiple simultaneous speakers. This foundational work, detailed in his most cited paper (over 200 citations), has become a standard platform in the field. Tsujino’s research addresses critical challenges in human-robot interaction, including robust sound source localization in dynamic settings, blind source separation with adaptive algorithms, and the suppression of a robot’s own motor noise. His innovative approaches, such as integrating audio-visual information and using missing feature theory, have significantly improved speech recognition accuracy for robots operating in cluttered spaces. With multiple papers garnering over 60 citations each, Tsujino’s work has profoundly shaped how robots listen, enabling more natural and effective communication between humans and machines.
Research Focus
Key Achievements
Top Papers
- 1
- 2Intelligent sound source localization for dynamic environments96 citations · 2009
- 3An open source software system for robot audition HARK and its evaluation70 citations · 2008
- 4
- 5
- 6
- 7Ego noise suppression of a robot using template subtraction52 citations · 2009
- 8
- 9Adaptive step-size parameter control for real-world blind source separation46 citations · 2008
- 10