When these particles touch, they don't exchange data — they exchange force. Each tentacle stores mechanical energy and releases it on contact, pushing neighboring particles away. No sensors. No chips. No batteries. Separation is built into the structure itself.
A LEGO brick is not smart. It doesn’t compute. It doesn’t plug in. It just fits.
A team of Georgia Tech researchers has applied that logic to robotics.
Bolei Deng, an assistant professor in Georgia Tech’s Daniel Guggenheim School of Aerospace Engineering, and Xinyi Yang, an aerospace engineering PhD student, build swarms of tiny robotic particles that latch, release, and reorganize without a single electronic component. No sensors, no processors, and no code.
The research recently appeared on the cover of Advanced Intelligent Systems.
Deng didn’t invent the idea. Novelist Kurt Vonnegut imagined self-organizing machines more than 60 years ago. [See sidebar.]
Deng’s goal was to turn that science fiction into fact.
“Instead of using a central controller, our particles’ behavior is governed by their mechanical design and how they interact with one another,” Deng said.
Traditionally, building smarter robots means adding complexity: more hardware, more processors, more code. Deng and Yang stripped all of that away.
What remains is mechanics.
Yang calls it “mechanical intelligence.” Instead of using sensors or a central brain, each particle is designed to let its shape do the “thinking.”
“The intelligence isn’t programmed in — it’s built in,” Yang explained. “Change the geometry, and you change what the swarm does.”
When the particles feel vibration, they respond automatically. Mix different shapes together, and the group starts to move like a flock of birds or a colony of ants.
“Each unit can be very dumb and follow simple rules. But when you combine enough of them, a sort of intelligence begins to emerge.” — Bolei Deng
Intelligence Without a Brain
In three dimensions, the same rule holds: shape controls behavior. Each particle’s geometry determines how it pushes, locks, or releases its neighbors. The result is coordinated movement without signals or code — interaction written directly into form.
Each particle in the swarm is identical — and completely useless on its own.
“Each unit can be very dumb and follow simple rules,” Deng said. “But when you combine enough of them, a sort of intelligence begins to emerge.”
Each particle has flexible arms spaced evenly around its body. When two particles meet, the arms bend and latch, storing tension like a compressed spring. An external vibration releases that stored tension. The arms snap open, the particles push apart, and the swarm spreads.
How far they spread, and how quickly, depends on how the arms are built. Change the curvature, and they hold on longer. Make them stiffer, and they release faster. Each particle follows the same simple mechanical rules: bend, latch, release.
Tiny Particles. Massive Stakes.
A single vibration sets the system in motion. Particles break apart in a defined sequence, each interaction triggering the next. No central control. The order of disassembly is pre-programmed in how the particles physically connect. .
The particles can be built at dramatically different scales — from the width of a human hair all the way up to 1.5 inches in size.
At their smallest, particles can enter the bloodstream. Doctors could place a compact swarm inside the vascular system and activate it with ultrasound. The vibration releases the stored tension in the arms. The particles spread outward and enter vessels a single robot cannot reach.
Deng envisions swarms delivering cancer drugs directly to hard-to-reach tumors while sparing healthy tissue. The approach targets diseased cells without compromising the rest of the body.
The swarm may also be able to map blood vessels, extending beyond the reach of today’s medical imaging tools.
“These particles could explore vessels no camera or catheter can reach,” Yang said. “You send the vibration, and they spread into parts of the body we can’t otherwise see.”
The same approach could work beyond the body. In space, even small fixes require astronauts to suit up for risky spacewalks, and radiation degrades electronics.
The particles could be launched as a compact cluster, land on a surface, and then be released with vibration. They spread out, move around obstacles, and reconfigure without sending anyone outside.
Because their behavior is built into their structure — and not electronics — the swarm could operate in radiation and temperature extremes that disable conventional robots.
“In space, once you build something, you need an astronaut or a robot to change it,” Deng said. “In our system, you just send the vibration.”
What Comes Next
Deng and Yang have shown that mechanics alone can move a swarm. Now they are pushing this idea further.
They are building structures whose joints respond to different vibrations. One pulse unlocks one joint. Another pulse releases a different section. The structure doesn’t just move — it rearranges itself. No processor chooses what shifts. The design does.
“We’re still just scratching the surface of what’s possible when you let the design do the work,” Yang said.
It’s the same LEGO logic the research team started with. No electricity required.
About Georgia Tech Commercialization
Georgia Tech Commercialization provides a foundation for faculty seeking to translate the Institute's leading-edge research into real-world applications. Commercialization encompasses four pivotal units: CREATE-X, VentureLab, Quadrant-i, and Technology Licensing. These units empower students and faculty to launch startups, provide comprehensive commercialization support, manage intellectual property, and facilitate the transformation of research into viable businesses. The Office of Commercialization’s mission is to provide world-class commercialization services, catalyzing research and innovation to improve the human condition, and enhancing Georgia Tech's position as a leader in technology and entrepreneurial impact.
First It Was Fiction. Now It’s Physics.
For more than 20 years, computer vision pioneer William Freeman challenged his MIT students with an unsolved problem. No one cracked it.
It started with a science fiction novel. In Cat’s Cradle, published in 1963, Kurt Vonnegut imagined “Ice-nine,” a substance that reorganizes matter through touch alone. Freeman set out to build a real-world version of Ice-nine. Colleagues told him computers just weren’t fast enough to design it yet. Come back in 20 years. He kept asking anyway.
Then Deng arrived.
Bolei Deng came to MIT as a mechanical engineering postdoc, not a computer scientist. He approached the challenge from a different direction — and it worked. Freeman is now a co-author on Deng’s paper.
A challenge that stumped a generation of computer scientists was solved by someone trained in a different discipline entirely.
Paradigm Shifters
At Georgia Tech, cutting-edge research is a constant — it fuels discovery, challenges norms, and inspires progress every day. And occasionally, technologies emerge with the power to fundamentally change life for the better.
These breakthroughs transform how we live, how we work, and even how we heal. Many years and minds in the making, these technologies are the culmination of collaboration, persistence, and ingenuity. These are the paradigm shifters.
Discover these extraordinary innovations as they begin to make their mark on the world. The future is taking shape right here at Georgia Tech.
Sheepdogs Reveal a Better Way to Guide Robot Swarms 
Engineering A Robot That Can Jump 10 Feet High – Without Legs 
Why Mosquitoes Swarm Your Head: They’re Following Signals, Not Each Other