3D printing robots use AI machine learning to create shock-absorbing shapes that humans just can't achieve

In a laboratory at Boston University's School of Engineering, a robotic arm drops small plastic objects into a box placed flush against the floor to catch them as they fall. These tiny structures, feather-light cylindrical objects less than an inch tall, are packed into boxes one after another. Some are red, others blue, purple, green, and black.

Each object is the result of an experiment in robotic autonomy, where the robot is learning to explore and create the most efficient energy-absorbing shape it has ever seen.

To achieve this, the robot creates a small plastic structure with a 3D printer, records its shape and size, and transfers it to a flat metal surface before an adult Arabian horse quarters it. crush with the same pressure as standing on the ground.

The robot then measures how much energy the structure absorbed, how its shape changes after being crushed, and records all the details in a massive database. It then drops the crushed object into a box, wipes the metal plate clean, and prepares the next piece to be printed and tested.

The robot is only slightly different from its predecessor, and its design and dimensions are adjusted by the robot's computer algorithms based on all previous experiments, using a technique called Bayesian optimization. With each experiment, the 3D structure becomes better at absorbing the impact of being crushed.

These experiments were made possible thanks to ENG Associate Professor of Mechanical Engineering Keith Brown and his team at KABlab. The robot, named MAMA BEAR (an abbreviation for its long formal title, Mechanics Bayesian Experimental Autonomous Researcher for Additive Manufacturing Architecture), has evolved since it was first conceptualized by Brown and his lab in 2018.

By 2021, the lab had begun work to create the machine's shape that would absorb the most energy, a property known as mechanical energy absorption efficiency. The current iteration has been running continuously for more than three years, with more than 25,000 3D-printed structures packed into dozens of boxes.

Why are there so many shapes? Things that can efficiently absorb energy have countless uses, from cushioning in delicate electronic devices shipped around the world to knee pads and wrist guards for athletes. there is.

“We can pull data from this data library to, say, make better car bumpers or better packaging equipment,” Brown said.

To function ideally, a structure needs to strike a perfect balance: it can't be so strong that it damages whatever it's supposed to protect, but it needs to be strong enough to absorb impacts.

Before MAMA BEAR, the best structures ever observed had an energy absorption efficiency of about 71 percent, Brown says. But on a chilly January afternoon in 2023, Brown's lab witnessed the robot reach 75 percent efficiency, shattering the known record. The results were published below. Nature Communications.

“When we started, we didn't know if we'd be able to create this record-breaking shape,” says Kelsey Snapp, a doctoral student in Brown's lab who oversees Mama Bear. “Slowly but surely, we made some incremental progress and made some breakthroughs.”

The record-breaking structure looks like something the researchers never expected: It has four points shaped like thin petals, and is taller and narrower than earlier designs.

“We have a ton of mechanical data here and we're excited to use it to learn lessons about design in general,” Brown says.

Their vast amount of data is already being applied to real life for the first time, helping design new helmet padding for U.S. Army soldiers.

Brown, Snapp, and project collaborator Emily Whiting, an associate professor of computer science in the BU College of Arts and Sciences, are working with the U.S. Army to create helmets with patent-pending padding that are both comfortable and protected from impact. A field test was conducted to confirm that protection can be achieved. . The 3D construction used in the padding differs from record-breaking items, with a softer center and lower profile for increased comfort.

MAMA BEAR isn't Brown's only autonomous research robot: His lab has other “BEAR” robots that perform a variety of tasks, including Nano BEAR, which uses a technique called atomic force microscopy to study the behavior of materials at the molecular level.

Brown is also working with Jörg Werner, an assistant professor of mechanical engineering at ENG, on another system called PANDA (short for Polymer Analysis and Discovery Array)-BEAR to test thousands of thin polymer materials to find the best ones for batteries.

“These are all robots doing research,” Brown says. “The idea is to use a combination of machine learning and automation to greatly increase the speed of research.”

“It's not just faster,” Snapp adds, “it can do things that it normally couldn't do – achieve structures and goals that would have been too costly and time consuming to achieve otherwise.” He has worked closely with MAMA BEAR since the experiment began in 2021, giving the robot the ability to see, known as machine vision, and the ability to clean its own test plates.

KABlab wants to further demonstrate the importance of autonomous research. Brown wants to continue collaborating with scientists from different disciplines who need to test a huge number of structures and solutions. Even though they've already broken the record, “we don't know if we've reached maximum efficiency,” Brown said, meaning it could be possible to break the record again.

So while MAMA BEAR continues to operate and push its boundaries even further, Brown and his team consider what other uses the database could be made of. He also unraveled more than 25,000 pieces and reloaded them into his 3D printer, exploring ways to recycle the materials and use them for further experiments.

“We will continue to study this system because, like many other material properties, mechanical efficiency can only be accurately measured experimentally,” says Brown. as soon as possible. ”

This article is reprinted with permission from Boston University. Read the original article here.