Subtitles section Play video Print subtitles We're here a Cornell University. And this is Carson Peterson's love, the collective embodied intelligence love. So in general, throughout all the projects in this lab, it's about creating large collectives of robots and minimizing ah, the amount of manufacturing that has school into into it so that weaken manufacturing large men robots. One of the projects is the self sleeve. We have a set of resistive sensors on a sleeve. Each of these black lines on the self censor are made up of a carbon black slash uncle flex gel, where they're electrically conductive and when you inflate it around an object. Ah, you can map out the shape of the object by measuring the amount of time that it takes for one of the lines of resistance to stop showing a change of resistance. The end goal of this leave is integrated into vineyards where you could have a self robot are, ah, self censor. Uh, go up to these grapes the grape clusters and inflated around the grapes without damaging them, and then be able to reconstruct the geometry and get other characters six out of it like, for example, its rightness. So you could also incorporate other types of sensors into this. So what bits of kind of harbor and software using for this? I mean, is it fairly stands toe. So I guess one of the main takeaways of this self censor is that it's relatively easy to make. Um, it's basically just a couple of layers of silicone and then one layer of, uh, carbon black, uh, and equal Flex Joe. And then you hook it up to the Arduino and you inflate it and you can get some data out of it. It's a pretty simple set up, so the idea would be that people can manufacture these sensors pretty easily. Right here. You have a bunch of output lines, and basically you have, ah, and input in electricity on one side. And then the output and you're measuring, measuring the change in resistance by taking it out to an Arduino. And in the Arduino, you can map it out onto a plot where you can see ah, the change in resistance as a function of time. There's one or a green. Oh, and then there's two multiplexes here to get the readings out, and these air, the one is to inflate the sleeve, and then the other is to vacuum hero. You got to compress it down. Yeah, So this one had needs Ah! Ah, Higher pressure toe. Actually be able to inflate the phone. Is this stuff three D printed? Uh, no, this is all molded. So it has a good amount of steps in the fabrication process. So basically, it's several layers of ICO flex type of silicone molded around a foam and foam gives you a given process. 50% brasi, like 50% open space for air to flow through it. That creates an amount of fluid resistance and causes a traveling with through the robot. Right now, the demo is yeah, for movement. And basically it's inching forward by having a traveling wave across the robot. But what we want to show mainly is that you can have a single inlet and then have a sequence of actuacion sze enabled by the flu dick resistance from the foam inside the software. Uh, we have some kind of leak, so use the compress to build up some pressure on your period. Yeah, it's hypnotic. Yeah, e guess something that we can see right here is, uh, I'll turn this off. So what we're trying to show right here is that you can have a single inlet of pressurization and then using the foam inside, um, with a 50% brasi of 50% amount of open space, Then you can create in mount an amount of fluid resistance. And then with that, you can use it to create a sequence of actuacion Sze. The neat thing about this is that it's embedded into the design of it, right? So the foam inside creates that amount of fluid of resistance that allows for the delay between the first leg and then and second line on the third. Like I guess, because it's their air is pretty invested. Ah has a very low vis viscosity, so you don't see the difference as much as if it was being inflated by water. This was an initial project. And then the next step, uh, that we're doing is creating a ah program that uses thes electrical circuit analogies to automatically design the structure of the foam inside. So if you let's say I have a given prosit e at the beginning and then a lower prosit e at the end, then I'll have a change in the amount of delay that certain sections of the phone give. And then I could have a programs even more control over the sequence. Evacuations that I have we have. He's, uh, fluid. Ah, slash electrical circuit analogies where you can model each of these segments each component of the robot as a politically resistive and a flu tickly capacitive region. And if I assign a certain value to that, then I can map out the amount of time that it would take the fluid to reach that part of the foam. If I send an impulse of pressure of air or water, the final blow of this would be to use that Ah, automate automatic design synthesis spurred him to say Okay, well, we have this robot. We have this general shape for the robot, and we want ah inlet right here and some, uh, outlets of pressurization at these points. And we want the sequence evacuation. And, um, and that could be a first swimmer or crawling robot. And by using, um, foam structure and a single inlet pressurization source, Then you could have this sequence evacuations for the software. What single wave We're working on more complex implementations. And finally, it can even perform something called Concertina Locomotion. And so, Ethan, that down to the hex engine, the control system, which is around the side here we come around the side.