A robot can lift a heavy box but still struggle to pick up a tiny wire without damaging it. Strength is not the problem. The challenge is knowing exactly how much pressure to apply.
A newly developed artificial skin could help solve this problem. It gives robots thousands of closely spaced touch sensors, allowing them to detect fine details, handle fragile objects and respond quickly to potentially damaging contact.
How the Artificial Skin Works
Researchers at the Southern University of Science and Technology in Shenzhen, China, developed a flexible sensing system containing more than 10,000 individual sensing units.
Their findings were published on October 16, 2026, in Nature Machine Intelligence.
The technology, called iontronic skin, uses electrically responsive materials to detect pressure. Each touch produces a measurable signal.
The important advance is how those signals are processed.
Conventional robotic sensor arrays often read their sensors one after another. As the number of sensors increases, that process becomes slower and generates substantial data traffic.
The new system identifies sensor signals using different frequencies, allowing many signals to be processed in parallel. It operates at up to 100 frames per second and reduces the amount of transmitted data by hundreds of times.
That means a robot can gather detailed touch information without overwhelming its electronic processing system.
What Changes for Robots?
The researchers demonstrated a robotic gripper handling a pencil lead just 0.5 mm thick and a soft solder wire measuring only 0.2 mm in diameter.
The artificial skin also detected fine surface features as small as 0.3 mm.
More importantly, it could recognize concentrated pressure associated with potentially harmful contact and trigger protective responses. This resembles a basic biological reflex, although the robot is not experiencing pain.
These capabilities could matter in electronics manufacturing, where robots must handle delicate components and tiny connectors. They could also improve laboratory automation and human-robot interaction.
Today, many industrial robots rely heavily on cameras, position measurements and force sensors. A dense layer of artificial skin would provide something different: information about where contact occurs and how pressure is distributed.
The Bigger Breakthrough Is Data Processing
Adding more sensors is relatively straightforward compared with processing thousands of signals quickly enough to control a moving robot.
The researchers addressed both problems together.
Their system combines dense sensing with a layered processing approach inspired by the human nervous system. This could eventually allow robots to respond to local contact without requiring every raw sensor measurement to be processed centrally.
What Happens Next?
The demonstrations remain laboratory experiments.
The next challenges include making the skin durable, affordable, easy to manufacture and reliable under repeated physical contact.
The researchers have not demonstrated that the technology can support human-level touch or general-purpose manipulation in real industrial environments.
Nevertheless, the work addresses an important obstacle in robotics: machines need more than intelligence and powerful motors. They also need reliable physical feedback about the world they touch.
Research source: Zhibin Li and colleagues, Large-scale iontronic skins with neural-hierarchical processing architecture, Nature Machine Intelligence, October 16, 2026. Original research paper.
Sam Brown holds a BSc in Electronic Systems Engineering from University of Regina, a Masters from the University of Alberta and a PhD from the University of Toronto in Electrical Engineering. His reporting interests include Physics and Artificial Intelligence.