These 3D-Printed Microrobots Use the "Cheerio Effect" as a Tool for Intelligent Swarming Behavior

Researchers at the Max Planck Institute for Intelligent Systems, alongside colleagues from Cornell University and Shanghai Jiao Tong University, have shown off the creation of swarming 3D-printed microrobots designed for rapid reconfiguration — including the ability to "dance the boogie."

The microrobots design takes advantage of what scientists call the "cheerio effect," named after the cereal of the same name: particles, or Cheerios, floating on the surface of a fluid will tend to drift towards one another if they pass close by, bending the surface of the water to eventually collide — while other particles may cause the water to raise and repel nearby objects. Combined with magnetism, that's enough to induce and control swarming behavior.

"Depending on how we change the magnetic fields, the discs behave in a different way," Gaurav Gardi, co-lead author of the paper, explains. "We are tuning one force and then another until we get the movement we want. If we rotate the magnetic field within the coils too vigorously, the force which is causing the water to move around is too strong and the discs move away from each other. If we rotate too slow, then the cheerio effect which attracts the particles is too strong. We need to find the balance between the three."

In a demonstration of the compact microrobots, the team steers them through a maze in which their collective behavior is key to reaching the end — switching into single file to pass through a narrow gap, for example, or forming a clump to push a plastic ball towards a goal. Another demonstration uses the microrobots to drive gears, while yet another lines the microrobots up like soldiers on parade. In all cases, the swarming behavior is directed through a computer control system based on an algorithm the team developed.

"Our vision is to develop a system that is even tinier, made of particles only one micrometer small," Gardi says of the team's future plans. "These collectives could potentially go inside the human body and navigate through complex environments to deliver drugs, for instance, to block or unblock passages, or to stimulate a hard-to-reach area."

"Robot collectives with robust transitions between locomotion behaviors are very rare. However, such versatile systems are advantageous to operate in complex environments," adds Metin Sitti, lead of the MPI-IS physical intelligence department and co-author of the paper.

"We are very happy we succeeded in developing such a robust and on-demand reconfigurable collective. We see our research as a blueprint for future biomedical applications, minimally invasive treatments, or environmental remediation."

The team's work has been published under open-access terms in the journal Nature Communications.