Robotic glove teaches your hand the basics of drawing
Whether it was after getting hooked on your first comic, taking a college art class, or even idly doodling on your math book instead of paying attention to your teacher, we’ve all experimented with drawing. Unless you’re one of the people that can actually do it well, you likely gave up and moved on, wondering how other humans can mix lines together to create something both recognizable and aesthetically pleasing. If you’re illustrationally-challenged, your salvation may lie not with humanity, but with robotics. A new robotic glove teaches you how to draw by becoming training your muscle memory.
Copenhagen Institute of Interaction Design student Saurabh Datta developed the glove as part of his thesis, initially as a way to learn to play the piano. If his human hands couldn’t learn, maybe some robot hands could teach them — and no, the robot hand doesn’t come from the Robot Devil, despite the startlingly similar way the idea was conceived. Called Teacher, the glove-like robot straps onto your hand and fingers, and guides you through specific gestures over and over. If you do it enough, your hand will learn how to do it through sheer muscle memory.
Obviously, this won’t teach you instinct or how to transfer something from your imagination to paper, but at the very least, the theory is that it’ll teach you basics — how to make aesthetically pleasing lines.
Now, it only took Datta a week to build the rig. It’s not exactly the teacher after which it’s named, but instead represents the way humans and robots can and do interact when working to achieve the same goal. Despite being presented with the potential to learn how to draw, Datta found that most participants didn’t like when the glove controlled the majority of the movement — they’d fight against the haptic feedback, and constantly readjust their hand within the contraption to find a more comfortable position. To fix the comfort issue, Datta recorded the fidgets made by the testers, and then adjusted the machine’s force feedback to account for them. In turn, this also helped the machine learn about the way humans naturally move.
Datta’s machines won’t suddenly help you create the best DeviantArt page the internet has ever known, but it’s essentially a proof-of-concept for machines doing our learning for us. You can check out the full project over here, including development diagrams and (long) demonstration videos.