Inside a lab at Boston University’s College of Engineering, a robot arm meticulously drops small, plastic objects into a box on the floor. These feather-light, cylindrical pieces, no bigger than an inch tall, are the results of an ongoing experiment in robot autonomy. Independently learning and adapting, the robot is on a mission to create the most efficient energy-absorbing shape ever conceived.

The process begins with the robot 3D printing a small plastic structure, recording its shape and size, and then placing it on a flat metal surface. The robot then crushes the structure with pressure equivalent to the weight of an adult Arabian horse standing on a quarter. It measures the energy absorbed by the structure, analyzes its deformation, and logs every detail into a vast database. The crushed object is then dropped into the box, and the robot prepares to print and test the next iteration.

This iterative process is driven by Bayesian optimization, where each new structure is slightly modified based on the data from previous experiments. Over time, these 3D-printed shapes improve in their ability to absorb impact. Since 2021, the lab has tasked the robot, affectionately named MAMA BEAR, with discovering the optimal shape for mechanical energy absorption efficiency. After running continuously for over three years, MAMA BEAR has produced more than 25,000 structures, filling dozens of boxes.

Why create so many shapes? Efficient energy absorption has countless applications, from cushioning delicate electronics during shipping to providing protection in sports equipment like knee pads and wrist guards. According to Brown, “You could draw from this library of data to make better bumpers in a car, or packaging equipment, for example.”

The structures must strike a balance between strength and flexibility: they need to be robust enough to absorb impact without damaging the objects they protect. Before MAMA BEAR, the best known structure was about 71% efficient at energy absorption. However, in January 2023, Brown’s lab achieved a new record of 75% efficiency, surpassing previous benchmarks. These groundbreaking results have been published in Nature Communications.

This innovative approach not only sets new standards in material science but also demonstrates the potential of autonomous robotics in optimizing complex processes and achieving unprecedented levels of efficiency.

By Impact Lab