Here’s a pretty impressive demonstration of a complex new speech recognition technology. The robot in this video clip is able to listen to three people speaking at the same exact moment, understand exactly what they each said, then process the data dynamically.
Be sure to have your speakers turned on for this video clip or you’ll miss how it works.http://video.google.com/videoplay?docid=-3386251792468042381
The Robot Audition Project is a collaborative effort by Kyoto University and Honda Research Institute Japan. The team has been working on developing these complex voice processing capabilities for several years now, but it’s only recently that they’ve got the system working so well. The application can not only recognize voices in real time, it can determine where they’re coming from using both auditory and visual cues, including face recognition technology.
The engineers who’ve designed the system have honed it enough now that they’ve even got it running on several different robotic platforms, including the Robovie R2 (seen in the video) and the well-known Honda Asimo.
Perhaps once they get these robots working in the real world, they’ll finally get my order right at the Burger King.