http://www.youtube.com/watch?v=USCKIiwareU
Currently in academia, most labs will spend time in the beginning of their project to develop their own hardware and then develop the code and controls to make it work. Some people say this is similar to the situation with computers in the 1980's, when many people developed their own platforms and made only incremental progress with their research, in large part due to the time intensive nature of that beginning phase. Things sped up once you had an operating system, like Windows or a Mac, and could jump right into developing higher level things, like Microsoft Word or an internet browser.
The idea behind the iCub robot is that you can start off with this hardware and dive right into the higher level research. So it is meant for things like machine learning, machine vision, and dexterous manipulation.
Right now the hand costs over $10,000 to machine out of aluminum. My job is to try and create it out of plastic, so you can 3D print it for much cheaper and the design would be much easier to modify. The only problem is that plastic is much weaker than metal so I have to play with things a bit.
Anyways, these are the models that I look at and alter all day in my CAD (Computer Aided Design) software. I've currently got a couple of fingers that I printed out, but broke since they were too flimsy.
No comments:
Post a Comment