Unleashing the Power of Swarms: Mind-Blowing Microrobots Self-organize into Mesmerizing Patterns

The approach may help inform how future microrobots could perform targeted drug delivery in which batches of microrobots transport and release a pharmaceutical into the human body.

The team’s paper, “Programmable Self-Organization of Heterogeneous Microrobot Collectives,” published June 5 in Proceedings of the National Academy of Sciences.

The lead author is Steven Ceron, Ph.D. ’22, who worked in the lab of the paper’s co-senior author Kirstin Petersen, an Aref and Manon Lahham assistant professor and faculty fellow in the Department of Electrical and Computer Engineering at Cornell Engineering.

Petersen’s Collective Embedded Intelligence Lab has been studying a variety of methods, from algorithms and classical control to physical intelligence, to convince large collectives of robots to behave intelligently, often taking advantage of the robots’ interactions with their own. environment and between them. However, this approach is extremely difficult when applied to micro-scale technologies, which are not large enough to accommodate on-board computing.

To meet this challenge, Ceron and Petersen partnered with the paper’s co-authors, Gaurav Gardi and Metin Sitti, of the Max Planck Institute for Intelligent Systems in Stuttgart, Germany. Gardi and Sitti specialize in the development of micro-scale systems driven by magnetic fields.

“The difficulty is how to enable useful behaviors in a swarm of robots that have no means of computing, sensing or communication,” Petersen said. “In our last paper, we demonstrated that by using a single global signal we could activate robots, in turn affecting their pairwise interactions to produce collective motion, contact- and non-contact-based object manipulation. We have now shown that we can expand that repertoire of behaviors further, simply by using different sizes of microrobots together, such that their pairwise interactions become asymmetric.”

The microrobots in this case are 3D-printed polymer discs, each about the width of a human hair, that have been sputter-coated with a thin layer of a ferromagnetic material and placed in a 1.5-meter pool of water. centimeters wide.

The researchers applied two orthogonal external oscillating magnetic fields and adjusted their amplitude and frequency, causing each microrobot to rotate on its central axis and generate its own fluxes. This movement in turn produced a series of magnetic, hydrodynamic and capillary forces.

“By changing the global magnetic field, we can change the relative magnitudes of those forces,” Petersen said. “And that changes the overall behavior of the swarm.”

By using microrobots of different sizes, the researchers demonstrated that they could control the level of self-organization of the swarm and how the microrobots assembled, dispersed, and moved. The researchers were able to: change the general shape of the swarm from circular to elliptical; force microrobots of similar size to group into subgroups; and adjust the spacing between individual microbots so that the swarm can collectively capture and expel foreign objects.

“The reason we’re always excited when systems are able to cage and expel is that you could, for example, drink a vial with little microrobots that are completely inert to your human body, have them enclose and carry medicine, and then take it to the right point on your body and let go,” Petersen said. “It’s not perfect object manipulation, but in the behavior of these micro-scale systems we’re starting to see a lot of parallels with more sophisticated robots despite their lack of computation, which is pretty exciting.”

Ceron and Petersen used a swarm oscillator model, or swarmalator, to accurately characterize how asymmetric interactions between disks of different sizes enabled their self-organization.

Now that the team has shown that the swarmalator fits such a complex system, they hope the model can also be used to predict new and never-before-seen swarming behaviors.

“With the swarmalator model, we can abstract the physical interactions and summarize them as phase interactions between swarm oscillators, which means we can apply this model, or others like it, to characterize behaviors in various swarms of microrobots,” said Ceron, currently a postdoctoral fellow at the Massachusetts Institute of Technology. “We can now develop and study collective behaviors of magnetic microrobots and possibly use the swarmalator model to predict behaviors that will be possible through future designs of these microrobots.”

“In the current study, we were programming the differences between the forces exerted across the size of the microrobots, but we still have a large parameter space to explore,” he said. “I hope this represents the first in a long line of studies in which we exploit heterogeneity in microrobot morphology to elicit more complex collective behaviors.”

The research was supported by the Max Planck Society, the National Science Foundation, the German Fulbright Scholarship, and the Packard Foundation Grant for Science and Engineering.