This robot isn't nearly as good as us at putting together IKEA furniture, but it's learning

Contributed by
Sep 29, 2015

It’s always fascinating to see a robot tasked with completing something humans find relatively easy (though your mileage may vary, admittedly). We now bring you: Skynet vs. IKEA.

Okay, not really Skynet — we’re actually talking about a robot built by researchers at the  Nanyang Technological University in Singapore. But it really was pitted against IKEA. Sadly, we’re still a long way away from robots capable of taking all those disparate parts and making pretty furniture. But the robot was able to use trial and error to figure out where it needed to input those annoying wooden pegs. Which, though it may seem trivial, represents some solid problem-solving skills from our future robo-overlords.

To work this robo-magic, the team integrated optical motion capture, motion/grasp planning and force control to achieve this “highly dexterous” task. If you’re looking for more intel on the technical process, they wrote a paper on the project called "A Framework for Fine Robotic Assembly.” You can read it here.

As The Verge notes, we already have robots capable of doing very specialized tasks very well (see: automotive component manufacturing). But what happens if you just give the robot all the parts and ask it to build the car? That’s where it gets tricky. The human brain is unique because it can use intuition and context to solve problems. Robots, not so much. So getting a robot to the point where it can figure out how to handle the pegs in an IKEA product is a small but significant move in the right direction.

Check out the IKEA ‘bot in action below and let us know what you think:

(Via The Verge)