Engineers in Japan have developed a soft robotic gripper that can genuinely "feel" the objects it touches — a breakthrough that could bring far more capable and delicate robots into factories, hospitals, and everyday life.
The device, called EleTac and developed at the Japan Advanced Institute of Science and Technology (JAIST), draws its design from one of nature's most versatile precision instruments: the tip of an elephant's trunk. The result is a two-fingered soft gripper that can sense shape, texture, and applied force without any external cameras or additional sensor layers bolted on.
Existing soft robotic grippers typically rely on cameras mounted outside the device, but this creates a fundamental problem: the moment the fingers wrap around an object, the camera's view is blocked. Researchers have tried to solve this by adding separate sensing layers to the grippers' fingers, but that approach makes the device heavier, more fragile, and more complex to manufacture.
EleTac takes a completely different approach. Instead of adding sensors to the outside, it turns the gripper itself into the sensor.
The device is made from a soft, rubber-like material with two flexible fingers. Rather than motors, EleTac uses vacuum pressure to close — air is sucked out through a pneumatic channel, causing the soft walls to bend inward and the fingers to gently pinch around an object. This softness is what gives it both precision and safety when handling delicate items.
The key innovation is hidden inside: a tiny camera positioned at the finger-end of the device, ringed by 18 LEDs that cycle through green, red, and blue light. As the gripper makes contact with an object, its soft outer skin deforms — and the internal camera captures exactly how those deformations change with every touch, shape, and applied force. In the researchers' own analogy, it is "like if we had eyes inside our fingertips."
Hundreds of contact images are fed to an AI algorithm trained to interpret the visual patterns: how much force is being applied, what shape the object is, and how far the fingers have bent. The system doesn't just sense that it's touching something — it can infer the geometry and mechanical properties of whatever it holds.
The JAIST team put EleTac through a series of experiments that showcased both its sensitivity and versatility. In one striking test, a robot used EleTac to find a pen buried beneath a thin layer of sand — completely invisible to any external camera. By pressing gently into the sand and analyzing how its own fingers deformed, the AI distinguished between touching loose sand and touching the rigid pen, estimated the pen's position, and lifted it cleanly from below.
In another test, EleTac wiped dishes with a sponge, adjusting its grip based on real-time tactile feedback to apply exactly the right amount of pressure — firm enough to clean, gentle enough not to break the dish.
Professor Van Anh Ho, the team's lead researcher, hopes EleTac will form the foundation of a new generation of robots capable of handling unpredictable, delicate tasks — sorting fragile produce, assisting in surgeries, or helping elderly individuals with daily life. The research was presented in July 2026 and represents one of the cleanest integrations of biological inspiration and AI in soft robotics to date.
By combining grasping and sensing into a single soft device, EleTac sidesteps the traditional trade-off between capability and complexity — and brings robots one step closer to something that has always defined human hands: the ability to understand the world through touch. Evolution spent millions of years perfecting the elephant's trunk. It took a team in Japan to show how much we still have to learn from it.
