Taiwan University Students Build Tour-Guide Robot
BobB-nw writes "A group of engineering students and their professor at National Taiwan University have built a robot that can map out the area it's operating in and offer guided tours. The robot, named 'Hsiao Mei,' uses laser mapping and GPS technology to navigate on its own, including around corners and obstacles such as tables and chairs. In the first public demonstration on Thursday, an engineering student with a wireless remote control first took the robot through the floor of a small museum on the university's campus so it could create a map for itself. After the initial run-through, the robot was able to make an unaided journey around the floor."
"On your left, you may observe a thirteenth century representa... GZHRRRZT... ANNIHILATE HUMANS! ANNIHILATE HUMANS!"
I've worked on a similar project, but we avoided GPS, since it doesn't work very well indoors.
What I want to know is if TFA is wrong, or if they're really using GPS. If it works fairly well, that would have made our lives a lot easier during my project.
Synthetic indoor GPS is commonly used for robotics.
http://michaelsmith.id.au
The problem with this sort of article is that it doesn't say how the thing actually works. They are trying to solve the 'SLAM' or Simultaneous Localization and Mapping problem. The best approaches have recently been particle filter based. See the work by Sebastian Thrun at Stanford or Montemerlo's dissertation at CMU. I would have to guess that they are using a similar approach.
But of course this has been done before. The discussed project looks very similar to Minerva.
The more people I meet, the better I like my dog.