A microrobot on a U.S. penny, showing its microscopic scale. (Image Credit: Michael Simari, University of Michigan)
Robots keep getting smaller, but until now there has been a stubborn lower limit. Once you go below a millimeter, autonomy usually disappears. Tiny machines can move, but only if they are tethered, magnetically guided, or steered from the outside. What researchers at the University of Pennsylvania and the University of Michigan are now reporting shows that this limit has been crossed.
In a study published in Science Robotics, the teams describe microscopic robots that are fully programmable, autonomous, and small enough to sit unnoticed on a fingertip. Each robot measures about 0.2 by 0.3 millimeters and is just 50 micrometers thick, roughly comparable to the size of many microorganisms.
Despite this scale, they can sense their surroundings, make simple decisions, and swim through liquid on their own, powered only by light.
A microrobot, fully integrated with sensors and an onboard computer, small enough to balance on the ridge of a fingerprint. (Image Credit: Marc Miskin, University of Pennsylvania)
“We’ve made autonomous robots 10,000 times smaller,” Marc Miskin, assistant professor of electrical and systems engineering at Penn and senior author of the work, said in the press release. “That opens up an entirely new scale for programmable robots.”
The difficulty, researchers explain, is not just miniaturization but physics itself. At human scales, motion is governed by gravity and inertia. At microscopic scales, those forces become negligible, and drag and viscosity dominate. “If you are small enough, pushing on water is like pushing through tar,” Miskin says. Strategies that work for larger robots, such as limbs or moving joints, tend to break or fail entirely when shrunk to this size.
Instead of fighting those constraints, the team designed a propulsion system that works with them. The robots do not flex or paddle. They generate an electric field that nudges ions in the surrounding liquid. Those ions then push nearby water molecules, effectively creating a tiny flow that carries the robot forward. “It’s as if the robot is in a moving river,” Miskin explains, “but the robot is also causing the river to move.”
A projected timelapse showing tracer particle trajectories around a microscopic robot consisting of three motors tied together.. (Image Credit: Lucas Hanson and William Reinhardt, University of Pennsylvania)
Because this system has no moving parts, the robots are remarkably durable. They can be transferred repeatedly between samples using a micropipette without damage and can keep swimming for months after being charged by an LED light. They reach speeds of about one body length per second and can be programmed to move in complex paths or coordinated groups.
Autonomy at this scale, however, requires more than motion. Each robot also carries a microscopic computer, memory, sensors, and solar panels, all squeezed onto a chip smaller than a grain of salt. That challenge was tackled by the University of Michigan team led by David Blaauw, whose lab has previously built record-setting sub-millimeter computers.
“The key challenge for the electronics is that the solar panels are tiny and produce only 75 nanowatts of power,” Blaauw says. “That is over 100,000 times less power than what a smart watch consumes.” To make that workable, the team designed circuits that operate at extremely low voltages and consume more than a thousand times less power than conventional designs.
The microrobot is equipped with a complete onboard computer, enabling it to receive and follow instructions autonomously. (Image Credit: Miskin Lab, Penn Engineering; Blaauw Lab, University of Michigan)
Space was just as limiting as energy. “We had to totally rethink the computer program instructions,” Blaauw says, explaining that tasks normally requiring many instructions were condensed into a single, specialized instruction so they could fit into the robot’s tiny memory.
The result, according to the researchers, is the first sub-millimeter robot with onboard sensing, computation, and autonomous control. Each robot can sense temperature differences as small as a third of a degree Celsius and respond by changing direction. When placed in liquid environments, this lets robots move towards areas of increasing temperature.
Communicating that information back to researchers required another unconventional solution. With no room for radios or antennas, the robots encode data in their movements. “To report out their temperature measurements, we designed a special computer instruction that encodes a value in the wiggles of a little dance the robot performs,” Blaauw says. “It’s very similar to how honey bees communicate with each other.”
The robots are both powered and programmed by light pulses, and each one has a unique address. This means different robots in the same group can receive different instructions and potentially perform different roles in a shared task.
Manufacturing also plays a key role in why this advance matters. The robots are built using standard CMOS fabrication, the same process used to make computer processors. That allows large numbers of robots to be produced on a single sheet and keeps costs low, with estimates as low as a penny per robot.
For now, these machines are simple, slow, and limited in what they can sense. But the researchers see them as a foundation rather than a finished product. Future versions could carry additional sensors, store more complex programs, move faster, or operate in more challenging environments.
“This is really just the first chapter,” Miskin says. “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.”