System capable of rapid transitions between behaviors – News Artificial intelligence and robotics

Researchers from the Max Planck Institute for Intelligent Systems (MPI-IS), Cornell University and Shanghai Jiao Tong University have developed collectives of microrobots capable of moving in any desired formation. Miniature particles are able to reconfigure their swarm behavior quickly and robustly. Floating on the surface of the water, the versatile microrobotic discs can spin in circles, boogie, cluster in a clump, spread out like gas, or form a straight line like beads on a string.

Each robot is slightly larger than the width of a hair. They are 3D printed using a polymer and then coated with a thin top layer of cobalt. Thanks to the metal, the microrobots become miniature magnets. Meanwhile, coils of wire that create a magnetic field when electricity passes through them surround the facility. The magnetic field allows the particles to be precisely directed around a centimeter wide pool of water. When forming a line, for example, researchers can move the robots so that they “write” letters in the water. The research project by Gaurav Gardi and Prof. Metin Sitti from MPI-IS, Steven Ceron and Prof. Kirstin Petersen from Cornell University and Prof. Wendong Wang from Shanghai Jiao Tong University titled “Microrobot Collectives with Reconfigurable Morphologies , Behaviors, and Functions” was Posted in Nature Communication April 26, 2022.

Collective behavior emerges from interactions between robots

Collective behavior and swarm patterns are found everywhere in nature. A flock of birds exhibits swarm behavior, just like a school of fish. Bots can also be programmed to act in swarms – and have been seen to do so quite visibly. A tech company recently presented a drone light show that earned the company a Guinness World Record by programming several hundred drones and flying them side by side, creating stunning patterns in the night sky. Each drone in this swarm was equipped with computing power directing it in every possible direction. But what if the single particle is so small that computation is not an option? When a robot is only 300 micrometers wide, it cannot be programmed with an algorithm.

Three different forces are at play to compensate for the lack of calculation. One is the magnetic force. Two magnets with opposite poles attract each other. Two identical poles repel each other. The second force is the fluid environment; water around the discs. When particles swim in a whirlpool of water, they displace the water and affect other surrounding particles in the system. The speed of the vortex and its magnitude determine how the particles interact. Third, if two particles are floating next to each other, they tend to drift towards each other: they bend the surface of the water in such a way that they slowly come together. Scientists and grain lovers call it the cheerio effect: if you float two cheerios on milk, they will soon bump into each other. On the other hand, this effect can also cause two things to repel each other (try a bobby pin and a cheerio).

Three forces enable reconfigurability

Scientists use the three forces to create a collective, coordinated movement pattern for several dozen microrobots in a single system. A video (https://youtu.be/q91AWmTBzG8) shows how scientists steer robots through a course, displaying the formation that best suits the obstacle course, such as when entering a narrow passage the microrobots line up in single file and disperse again when they go out. Scientists can also make the robots dance, alone or in pairs. Additionally, they show how they put a small plastic ball in the water tank and then aggregated the robots into a clump to push the floating ball. They can place the tiny particles inside two gears and move the particles in such a way as to make the two gears spin. A more ordered pattern is also possible, each particle keeping an identical distance with its neighbor. All of these different modes and formations of locomotion are achieved through external computation: an algorithm is programmed to create a rotating or oscillating magnetic field that triggers the desired movement and reconfigurability.

“Depending on how we change the magnetic fields, the discs behave differently. We grant one force, then another until we get the desired movement. If we spin the magnetic field inside the coils too vigorously, the force that is moving the water is too strong and the disks move away from each other. If we rotate too slowly, then the cheerio effect that attracts the particles is too strong. We have to find the balance between the three”, explains Gaurav Gardi. He holds a doctorate. student in the Department of Physical Intelligence at MPI-IS and one of the two main authors of the publication along with Steven Ceron of Cornell University.

A model for future biomedical and environmental applications

The future scenario for such microrobotic collectives is to go even smaller. “Our vision is to develop an even smaller system, made up of particles as small as one micron. These collectives could potentially get inside the human body and navigate complex environments to deliver drugs, for example, to block or unblock passageways, or to stimulate a hard-to-reach area,” says Gardi.

“Robot collectives with robust transitions between locomotion behaviors are very rare. However, such versatile systems are advantageous for operating in complex environments. We are very happy to have succeeded in developing such a robust and reconfigurable collective on demand. We see our research as a model for future biomedical applications, minimally invasive treatments or environmental remediation,” adds Metin Sitti, who heads the physical intelligence department and is a pioneer in the field of robotics. small scale and physical intelligence.

We wish to thank the writer of this article for this awesome web content

System capable of rapid transitions between behaviors – News Artificial intelligence and robotics


Explore our social media profiles and other related pageshttps://www.ai-magazine.com/related-pages/