MIT’s Bio-Inspired Robots

Watch this video to hear MIT Associate Professor Sangbae Kim’s students describe his biomimetics research. They explore why certain physical traits have evolved in animals in the natural world, then extract those useful principles, which are applicable to robotic systems, to solve such challenges as manipulation and locomotion in novel and interesting ways.



Transcript

00:00:00 [MUSIC PLAYING] SANGBAE KIM: This class is called the Bio-Inspired Robotics, but we actually teach the differences between biology and robotics. JESS HAN: The duality of Bio-Inspired Robotics really emphasizes that having a reference for something that works, and has been kind of vetted through natural selection, is really valuable, in terms of coming up with an idea

00:00:32 that you might not be able to otherwise. ROBERTO BOLLI: The idea is not necessarily to copy what we see in nature verbatim, but to understand why certain designs have evolved in nature. SANGBAE KIM: I think the ultimate goal of my class is actually to learn how human works, because it's impossible to build the machine that can help us without understand how we work. KENDRICK CANCIO: So the assignment is essentially to choose a dynamic system

00:00:58 and try to extract the biological principles. STUDENTS: Whoa. SANGBAE KIM: I think this year is one of the best years. There was a gymnastics robot, chicken robot, jumping robot. There was a hula hoop robot and there's a throwing robot, and there's a punching robot. [BELL RINGING] ZACK FRANCIS: I lovingly call it Mike Tyson. And what we wanted to explore was how the mass distribution on the forearm and the upper arm

00:01:23 segments would actually affect the momentum output. ERINA YAMAGUCHI: Our project was a kangaroo bot, but it was basically a single degree of freedom jumping leg that would jump on a trampoline. KENDRICK CANCIO: Our project is called Suni-bot, based off of Suni Lee, who is a USA gymnast. The core component of the project is to understand how gymnasts swing up on the uneven bars. ROBERTO BOLLI: We built a small bipedal walker that was running on a boot, affectionately known

00:01:53 as the chicken-bot 9,000. We found the chicken head, made the walker 300% cuter. JESS HAN: Through our simulations and our hardware experiments, we discovered that when you swing up as a gymnast, there's two different phases. And so it is this very dynamic, fast paced action. ERINA YAMAGUCHI: Going through my everyday life I kind of noticed it more, where those speeds that I walked with changed how I walked. So it was kind of nice to see how that all aligned.

00:02:22 ZACK FRANCIS: I boxed for about seven-ish years. Now I know how to put all my mass. And it's in the tip of the boxing gloves, obviously. KENDRICK CANCIO: In the class, you're consciously doing this trade off of, do I really need a bio inspired system to solve this problem? In some cases, the answer is yes, right? But in some cases, the answer is no. SANGBAE KIM: You don't need to constrain yourself into the bio-mimicry field.

00:02:47 We're talking about inspired robotics, not bio-mimicry robotics. So if there's an easier way to do in robotics, just don't be shy about it. ROBERTO BOLLI: When you look at animals, I always wonder now, why are they walking the way that they are? Why are they moving the way that they are? Is that just coincidental, or is there some strategy that they're employing to minimize energy, or to be extra fast?

00:03:12 [MUSIC PLAYING]