Scientists Develop Shape Shifting Robots That Mimic Human Touch

Engineers at EPFL have unveiled a groundbreaking set of soft robots capable of simulating human touch in 16 distinct ways. Dubbed “Digits,” these modular, air-powered machines can morph, vibrate, stiffen, and deliver tactile sensations without any coding needed. It’s a leap forward for virtual reality immersion, rehabilitation, and human-robot interaction.

Pneumatic Shape Shifting for Tactile Realism

The magic lies in pneumatic control. Digits modules, made of flexible joints and rigid links, use pressurized air pouches to transform shape and texture dynamically. They can simulate firmness, movement, and even subtle vibrations, all critical for reproducing realistic human touch.

Two prototypes illustrate the potential: TangiGlove, an exoskeleton glove that provides stiffness and tactile cues, and TangiBall, a handheld module that morphs into multiple shapes, from cubes to spheres, while exerting physical feedback. Together, they demonstrate how soft robots can adapt to both open- and closed-chain interactions, expanding the scope of possible applications.

Soft Robots Bridging Reality And Therapy

Humans perceive touch through complex interactions: pressing, rubbing, gripping, sensing texture and pressure. Most existing haptic devices only scratch the surface. Digits bridge that gap by combining shape change, movement, and vibration in one system. Their modularity enables endless tactile configurations tailored to individual users or therapeutic goals.

Even better, EPFL’s platform is built on Feelix, an open-source robotics framework with an intuitive interface. It uses machine learning to translate user intent into haptic sensations, no programming required. This removes a major adoption barrier for developers in VR, AR, and rehab.

What’s Next?

The Reconfigurable Robotics Lab, led by Jamie Paik, plans to deploy Digits in rehabilitation clinics to study hand and muscle recovery over extended periods. Meanwhile, developers are exploring more complex modules, richer textures, nuanced stiffness, and integration into full-body interfaces for immersive virtual environments.

“We aim to redefine human-machine interaction through robots that transform shape, stiffness, and haptic feedback,” Paik explains. The goal is to create systems that work for all users, regardless of size, ability, or task demands.

The full study is detailed in Advanced Intelligent Systems. This breakthrough puts EPFL at the cutting edge of soft robotics, turning lifeless machines into tactile collaborators capable of humanlike touch.