Barely bigger than a grain of dust and cheaper than a penny, a new generation of microscopic robots can swim, sense their surroundings and run tiny computer programs -- all on their own.
Researchers at the University of Pennsylvania and the University of Michigan have unveiled what they say are the world's smallest fully programmable, autonomous robots. Each machine is roughly 0.2 by 0.3 by 0.05 millimeters, operating at the scale of many microorganisms and just visible to the naked eye.
Despite their size, the robots can move in complex patterns, travel in coordinated groups, detect temperature changes and adjust their paths in response. They are powered and controlled by light, can operate for months in water and cost about a cent apiece to make, according to the team.
Marc Miskin, an assistant professor of electrical and systems engineering at Penn and senior author on two companion studies, said the work pushes robotics into a new regime.
"We've made autonomous robots 10,000 times smaller," he said in a news release. "That opens up an entirely new scale for programmable robots."
The robots are described in two papers, one in Science Robotics that focuses on their integrated computing and sensing, and another in the Proceedings of the National Academy of Sciences that details how they move.
Microscale robots have long been a goal for scientists who imagine tiny machines that could travel through the body, monitor individual cells or assemble microscopic devices. But shrinking a robot is not as simple as shrinking a computer chip. At very small scales, forces like drag and viscosity dominate, making motion extremely difficult. For decades, that challenge has slowed progress.
Miskin's group tackled the motion problem by rethinking what it means for a robot to "swim" at the microscale. Instead of relying on moving parts like propellers or legs, the robots move the water around them.
In water at this scale, the drag is so intense that it is like trying to push through tar. The team's propulsion system sidesteps this by using electric fields. The robots generate a field that nudges charged particles, or ions, in the surrounding liquid. Those ions then push on nearby water molecules, creating enough force to move the robot forward. The result is a propulsion system with no moving mechanical parts, which makes the robots durable and able to swim for months.
At the same time, the Michigan team led by electrical and computer engineering professors David Blaauw and Dennis Sylvester had been pushing the limits of tiny computers, building record-setting sub-millimeter systems that can sense and compute using minuscule amounts of power.
"We saw that Penn Engineering's propulsion system and our tiny computers were just made for each other," Blaauw, a senior author of the Science Robotics study, said in the news release.
To fit a functional "brain" on each robot, Blaauw's group had to work under extreme constraints. The onboard computer runs on about 75 nanowatts of power -- roughly 100,000 times less than a smartwatch. To harvest even that small amount, solar panels cover most of the robot's surface.
The limited power and memory forced the researchers to redesign how the robot's software works.
"We had to totally rethink the computer program instructions, condensing what conventionally would require many instructions for propulsion control into a single, special instruction to help us shrink the program's length to fit in the robot's tiny memory," Blaauw added.
The robots are programmed and powered by pulses of light. Each one has a unique identifier, so a researcher can shine patterned light and give different instructions to different robots in the same drop of water. In principle, that means a swarm of robots could divide up a task, with each unit taking on a specific role.
In the Science Robotics study, the team reports that this first batch of robots carries temperature sensors that can detect changes as small as a third of a degree Celsius. In lab tests, the robots could move toward warmer regions or report temperature as a stand-in for cellular activity.
To send back information, the robots change the way they move. The researchers liken this to the "waggle dance" that honeybees use to communicate. By wiggling in specific patterns, the robots can encode simple messages about what they sense.
Looking ahead, the team imagines many potential uses. In medicine, fleets of such robots might one day monitor the health of individual cells, track how tissues respond to drugs or deliver therapies with extreme precision. In manufacturing, they could help build or inspect microscale devices that are too small for human tools to handle.
Future versions could store more complex programs, move faster, integrate additional sensors or operate in more challenging environments beyond simple lab liquids. The current robots already show that it is possible to combine sensing, computing and motion at a scale that was out of reach just a few years ago.
"This is really just the first chapter," added Miskin. "We've shown that you can put a brain, a sensor and a motor into something almost too small to see, and have it survive and work for months. Once you have that foundation, you can layer on all kinds of intelligence and functionality. It opens the door to a whole new future for robotics at the microscale."
The project brought together experts in robotics, microelectronics and materials from Penn and Michigan, with support from the National Science Foundation and several other agencies and foundations. Graduate students Maya Lassiter at Penn and Jungho Lee at Michigan are co-first authors on the work.
For now, the robots swim only in carefully prepared lab environments. But as the technology matures, the researchers say, it could reshape how scientists and engineers interact with the microscopic world -- not just observing it through a microscope, but sending in tiny machines to sense, think and act on their own.