CC BY-ND” width=”800″ height=”394″/>

This connection of springs is a new type of material that can change shape and learn new properties. Credit: Jonathan Hopkins, CC BY-ND

A new type of material can learn and improve its ability to handle unexpected forces thanks to a unique grid structure with connections of variable stiffness, such as described in a new article by my colleagues and me.

The new material is a kind of architectural material, which derives its properties mainly from the geometry and specific features of the design, rather than from the material from which it is made. Take Velcro fasteners such as Velcro, for example. It doesn’t matter if it’s made of cotton, plastic or any other fabric. As long as one side is a fabric with stiff hooks and the other side has fluffy loops, the material will have the tacky properties of Velcro.

My colleagues and I based the architecture of our new material on that of an artificial neural network – layers of interconnected nodes that learn to do tasks by changing how much importance or weight they place on each connection. We hypothesized that a mechanical grid with physical nodes could be trained to assume certain mechanical properties by adjusting the stiffness of each connection.

To find out whether a mechanical lattice could take on and maintain new properties, such as taking on a new shape or changing its directional strength, we started by building a computer model. We then selected a desired shape for the material and input forces and had a computer algorithm match the stresses of the joints so that the input forces would produce the desired shape. We did this training on 200 different lattice structures and found that a triangular lattice was best to achieve all shapes tested.

<img src="" alt="Een nieuw type materiaal, een mechanisch neuraal netwerk genaamd, kan leren en zijn fysieke eigenschappen veranderen om aanpasbare, stro" title="Gebouwde materialen – zoals dit 3D-rooster – ontlenen hun eigenschappen niet aan het materiaal waarvan ze zijn gemaakt, maar aan hun structuur. Krediet: Ryan Lee, CC BY-ND


Built materials – such as this 3D grid – do not derive their properties from the material from which they are made, but from their structure. Credit: Ryan Lee, CC BY-ND

Once the many compounds are tuned to accomplish a set of tasks, the material will continue to respond in the desired way. The training is – in a sense – remembered in the structure of the material itself.

We then built a prototype physical grid with adjustable electromechanical springs arranged in a triangular grid. The prototype is made of 6-inch joints and is about 2 feet long and 1½ feet wide. And it worked. When the grid and algorithm worked together, the material was able to learn and change shape in certain ways when subjected to different forces. We call this new material a mechanical neural network.

next to what living tissues, very few materials can learn to better cope with unexpected loads. Imagine an airplane wing that suddenly catches a gust of wind and is forced in an unexpected direction. The wing cannot change the design to be stronger in that direction.

The prototype lattice material we designed can adapt to changing or unfamiliar conditions. For example, in a wing, these changes can be the build-up of internal damage, changes in how the wing is attached to a vessel, or fluctuating external loads. Each time a wing made of a mechanical neural network experienced one of these scenarios, it was able to strengthen and soften its connections to maintain desirable properties such as directional strength. Over time, through successive adjustments by the algorithm, the wing takes on and maintains new properties, adding each behavior to the rest as a kind of muscle memory.

<img src="" alt="Een nieuw type materiaal, een mechanisch neuraal netwerk genaamd, kan leren en zijn fysieke eigenschappen veranderen om aanpasbare, stro" title="Het prototype is 2D, maar een 3D-versie van dit materiaal zou veel toepassingen kunnen hebben. Krediet: Jonathan Hopkins, CC BY-ND“/>

The prototype is 2D, but a 3D version of this material could have many uses. Credit: Jonathan Hopkins, CC BY-ND

Related Post

This type of material could have far-reaching applications for the longevity and efficiency of built structures. Not only could a wing made from a mechanical neural network material be stronger, it could also be trained to morph into shapes that maximize fuel efficiency in response to changing conditions around it.

Until now, our team has only worked with 2D grids. But using computer modeling, we predict that 3D grids would have much greater learning and adaptability. This increase is due to the fact that a 3D structure could have tens of times more connections, or springs, that do not intersect. However, the mechanisms we used in our first model are far too complex to support in a large 3D structure.

The material that my colleagues and I have created is a proof of concept and shows the potential of mechanical neural networks. But to bring this idea into the real world, you have to figure out how to make the individual pieces smaller and with precise properties of flex and tension.

We hope that new research in the fabrication of materials on a micron scalebut also to work on new materials with adjustable stiffnesswill lead to advances that will make powerful smart mechanical neural networks with micron-scale elements and dense 3D connections a ubiquitous reality in the near future.

AI material that learns behavior and adapts to changing circumstances

Provided by The Conversation

This article was republished from The conversation under a Creative Commons license. Read the original article.

Quote: Material called a mechanical neural network can learn and change its physical properties (2022, October 21) Retrieved October 21, 2022 from -physical.html

This document is copyrighted. Other than fair dealing for personal study or research, nothing may be reproduced without written permission. The content is provided for informational purposes only.