Skip to main content

23 November 2026

Scientists create world's first tactile robot inspired by elephant's trunk

The robotic fascimile has even more sensitivity to an object’s shape, size and texture than its natural counterpart.

Elephant

Scientists have made the world's first tactile robot inspired by an elephant’s trunk with a sense of touch, capable of fine motor control like handling grapes and picking up playing cards.

Based on one of the animal kingdom’s best designed appendages, the device designed at King’s College London, in collaboration with researchers at the Japan Advanced Institute of Science and Technology, uses ‘vision-based’ sensors in the trunk’s soft tip to give ‘EleTac’ even more sensitivity to an object’s shape, size and texture than its natural counterpart.

By taking forward this nature inspired design, the team hope to create the next generation of robotic grippers that can both sense and manipulate objects, for use in areas like manufacturing, search and rescue and even fossil digging.

The elephant’s trunk is... the perfect natural tactile sensor. By learning from this, we’ve created a robotic gripper that can handle objects safely and with enough tactile feedback to handle a whole range of shapes, weights and sizes, making it flexible, adaptable and useful.’

Xuyang Zhang, PhD student

Professor Shan Luo, Professor of Robotics and AI at King’s and author of the paper, said ‘Tactile robots, like EleTac, are robots with a human-like sense of touch which makes them much more dexterous than traditional grippers you may see in a factory. While the human brain inherently knows the difference between picking up a grape and a metal pipe and applies different levels of force, a robot does not. This can lead to issues when dealing with a range of objects of different shapes and sizes.'

Xuyang Zhang, a PhD student at the Robot Perception Lab led by Professor Luo at King’s working on the project, continued ‘The elephant’s trunk, which essentially acts as the poor-sighted animal’s hand, has a soft tip able to delicately handle a range of objects without damaging them while also exploring its surroundings via touch – the perfect natural tactile sensor. By learning from this, we’ve created a robotic gripper that can handle objects safely and with enough tactile feedback to handle a whole range of shapes, weights and sizes, making it flexible, adaptable and useful.’

The sensor in the trunk’s tip works by housing a camera in a soft, rubber-like silicone shell. As the gripper handles the object, said object imprints itself on the silicone and deforms it. The camera records these deformations in the shell and uses AI to estimate the size, shape and force needed to manipulate the object, feeding back this information to inform the gripper’s approach. 

Taking this ‘vision-based’ approach the camera can sense forces as small as 0.25N, roughly the force half a chocolate bar exerts on the earth when on the ground, and similar to what the 400,000 nerve endings in an elephant’s trunk can feel.

Current tactile prototypes often use several high accuracy/torque sensors that can cost upwards of £10,000, moulding them into the shape of a human hand to grip and manipulate objects. In comparison, the soft tip of EleTac’s shell costs 50 pence to produce and under £75 for the entire device, making it on average 99% cheaper than similar industry models.

This work builds on the team’s earlier SimTac, a simulation environment that applied lessons from some of nature’s most famous ‘sensors’, like cats’ paws and elephant trunks, to cut the time and cost needed to create next-generation tactile sensors and robots.

In this story

Shan Luo

Professor in Engineering