Dr. Károly Zsolnai-Fehér shows a 3D-printing method that designs objects by how they deform, not how they look. It orients fiber-like structures to make things collapse or stay rigid in specific ways, and works on cheap printers. Five examples: pliers, bendy plates, and knee-like structures.
Dear fellow scholars, this is two minute papers with Dr. Zsolnai. I would like to show you some results from the area of 3D printing, a topic which is I think a little overlooked and show you that works in this field are improving at an incredible pace. Now a common theme among research papers in this area is that they typically allow us to design objects and materials by thinking about how they should look. Lets see if this is really true by applying the second law of papers which says whatever you are thinking about. There is already a two Minute Papers episode on that. Let's see if it applies here. For instance, just prescribing a shape for 3D printing is old, old news. Here is a previous technique that is able to print auxetic materials. These are materials that we can start stretching and if we do, instead of thinning, they get fatter. We can also 3D print filigree patterns with ease. These are detailed thin patterns typically found in jewelry, fabrics and ornaments. And as you may imagine, crafting such motifs on objects would be incredibly laborious to do by hand. We can also prescribe an image and 3D print an object that will cast a caustic pattern that shows exactly that image. Beautiful. And printing textured 3D objects in a number of different ways is also possible. This is called hydrographic printing and is one of the most flamboyant ways of doing that. So what happens here? Well, we place a film in water, use a chemical activator, spray on it and shove the object in the water and oh yes, there we go. Note that these were all showcased in previous episodes of this series. So in 3D printing we typically design things by how they should look. Of course. How else would we be designing? Well, the authors of this crazy paper don't care about looks at all. Well, what else would they care about if not the looks? Get this. They care about how these objects deform. Yes, with this work we can design deformations and the algorithm will find out what the orientation of the fibers should be to create a prescribed effect. Ok, but what does this really mean? This means that we can now 3D print really cool fiber like microstructures that deform well from one direction. In other words, they can be smashed easily and flatten a great deal during that process. I bet there was a ton of fun to be had at the lab on this day. However, research is not only fun and joy. Look, if we turn this object around, ouch. This side is very rigid and resists deformations well. So there was probably a lot of injuries in the lab that day too. So clearly this is really cool. But of course, our question what is all this good for? Is this really just an interesting experiment or is this thing really useful? Well, let's see what this paper has to offer in five amazing experiments. Pliers. The jaws and the hand grips are supposed to be very rigid. Checkmark. However, there needs to be a joint between them to allow us to operate it. This joint needs to be deformable. And not any kind of deformable, but exactly the right kind of deformable to make sure it opens and closes properly. Loving this one. 3D printing pliers from fiber like structures. How cool is that? Structured plates. This shows that not all sides have to have the same properties. We can also print a material which has rigid and flexible parts on the same side a few inches apart, thus introducing interesting directional bending characteristics. For instance, this one shows a strong collapsing behavior and can grip our finger at the same time. Bendy plates. We can even design structures where one side absorbs deformations while the other one transfers it forward, bending the whole structure. Example number four. Seat like structures. The seat surface is designed to deform a little more to create a comfy sensation. But the rest of the seat has to be rigid to not collapse and last a long time. And finally, example number five. Knee like structures. These freely collapse in this direction to allow movement. However, they resist forces from any other direction. And these are really just some rudimentary examples of what this method can do. But the structures showcased here could be used in soft robotics, soft mechanisms, prosthetics, and even more areas. The main challenge of this work is creating an algorithm that can deal with these breaking patterns which make for an object that is nearly impossible to manufacture. However, this method can not only eliminate these, but but it can design structures that can be manufactured on low end 3D printers. And it also uses inexpensive materials to accomplish that. And hold on to your papers, because this work showcases a handcrafted technique to perform all this. Not a learning algorithm in sight. And there are two more things that I really liked in this paper. One is that these proposed structures collapse way better than this previous method. And not only the source code of this project is available, but it is available for you to try on one of the best websites on the entirety of the Internet. Shadertoy. So good. So I hope you now agree that the field of 3D printing research is improving at an incredible pace. And I hope that you also had some fun learning about it. What a time to be alive. This episode has been supported by Weights and Biases. In this post they show you how to use their tool to extract text with all kinds of sizes, shapes and orientations from your images. During my PhD studies I trained a ton of neural networks which were used in our experiments. However over time there was just too much data in our repositories and what I am looking for is not data but insight. And that's exactly how Weights and Biases helps you by organizing your experiments. It is used by more than 200 companies and research institutions including OpenAI, Toyota Research, GitHub and more. And get this Weights and Biases is free for all individuals, academics and open source projects. Make sure to Visit them through wnb.com papers or just click the link in the video description and you can get a free demo today. Our thanks to WeightsAndBiases for their long standing support and for helping us make better videos for you. Thanks for watching and for your generous support and I'll see you next time.
