This summer, I am working with an undergrad team at Columbia’s Robotics Lab in the computer science department. I am focusing on the mechanical engineering end of a very interesting project for the handicapped. I’ll be blogging about my experiences as the summer goes by.
The long term goal of the project is to create a robotic wheelchair and arm system (in the far future, possibly a two-arm system) for paraplegics, run via Brain Computer Interface (BCI) — basically, you can put on a lightweight BCI headgear and control the chair and arm. You may have seen a similar project on the front page of the Wall Street Journal recently, which featured the “BrainGate” BCI system. While the BrainGate system is a huge step, it has some flaws — most significantly it is invasive (you have to drill a hole in your skull and latch a sensor to your brain, which is not fun), while ours would be non-invasive with simple, lightweight headgear; the Columbia Robotics Lab already has a small system which works for this. Secondly, the WSJ article features a robotic arm which is heavy and costly.
This second aspect is where my work comes in. Our lab team’s short term goal is to create a mobile manipulator (a moving base and arm system) to battletest ideas for the wheelchair system. Our aim is to design and create a cheap, relatively high level of control arm able to operate within a domestic environment. This involves being able to reach objects on a desk, shelf, etc., being able to reach around obstacles, picking up or moving household items, and so on. We plan to have a at least a 6 degree of freedom (DOF) arm, though we will likely do 7 DOF.
For the past few weeks we have thoroughly researched material on other arms and mobile manipulators, in addition to series elastic actuators — joins witch allowing the arm to “give” (compliance). If it hits a person, or object, the arm should be able to fold back a bit by providing compliance at joints, that way it don’t hurt the person or damage the object. This danger of collision in the small confines of a home is one reason why we lightweight arms are preferable. While it’s possible, as in the BrainGate system, to program an arm to stop immediately upon impact, it’s very useful to have a built in, physical safety component. We are modeling the design partly off of the WAM arm (a common industrial and research arm, but very costly) and an extremely lightweight Stanford arm made of plywood (great paper here), though we have considered some other more creative ideas, such as this funky inflatable robotic arm or Festo’s bionic handling assistant (which looks like Doc Oc, something straight out of Spider Man 2), though we ultimately found such ideas impractical for our uses.
This week we’ve started using ProE to design our first prototype. Last week we also looked at methods of allowing for give, and will continue that as well. Very exciting.
Meanwhile, I’ve also been working part time at Scholastic, Inc. for The Alliance For Young Artists and Writers for their annual print publication, The Best Teen Writing of 2012 — past years’ Best Teen Writing are available on Amazon if you want to purchase them, and this year’s anthology will come out this fall. The Alliance hosts the most prestigious contest for teen writers and artists, the Scholastic Art and Writing Awards: 200,000 submissions of art and writing worldwide, 1,600 national winners, and over 500 in the writing category. The contest, which is reaching its 90th anniversary this year, is famous for picking up prestigious artists before they make a name for themselves-past winners include Truman Capote, Richard Avedon, Robert Redford, and Andy Warhol (though if you saw Men in Black 3 you know he’s really a secret agent). The Best Teen Writing is a collection of the top 50 of those top 500 in the nation — the best of the best.