Shape-shifting soft robots deliver real human touch with 16 powerful haptic modes

Developed by the Reconfigurable Robotics Lab at EPFL’s School of Engineering, Digits is modular in nature.

It uses flexible joints and rigid links activated by pressurized air pouches to change form and tactile output. Despite supporting 16 unique configurations, its underlying design remains simple.

Modular devices show real-world versatility

Two working prototypes, TangiGlove and TangiBall, demonstrate the system’s adaptability. The TangiGlove, an open-chain exoskeleton worn on the hand, delivers stiffness feedback and tactile cues.

The TangiBall, a hand-held closed-chain module, morphs into eight distinct shapes, including a cube and sphere, while providing similar feedback and vibration.

“Haptic, or tactile, interfaces can enhance virtual reality experiences by emulating real-world touch, and support rehabilitation through interactive systems,” said Serhat Demirtas, a Ph.D. student and first author of the study. “But there is a real need for more generalized reconfigurable designs and control methods.”

Bridging the gap in haptic realism

Human touch is inherently complex. Unlike vision or hearing, it requires interaction, such as grasping, pressing, or rubbing, to sense texture, temperature, or pressure.

Most existing haptic devices can only mimic one of these traits. Digits aims to overcome that limitation.

The system spans both major robotic categories: open-chain and closed-chain configurations.

This dual capability allows a wide range of movements and transformations. The modular design also ensures easy customization for different users and tasks.

Smart interaction, no coding required

To make the system user-friendly, the team expanded the open-source robotics platform Feelix. This lets users design custom haptic responses without writing code.

The platform uses machine learning to interpret how the Digits respond to touch and generate intuitive feedback in real time.

Digits also stands out by relying on pneumatic actuation, an area often overlooked in haptic tech. Air-powered systems offer fine control over shape and stiffness, helping simulate real-world tactile experiences with high precision.

Next steps in therapy and interaction

Led by Jamie Paik, the lab plans to test the Digits framework in rehabilitation settings.

They aim to evaluate its long-term use for motor recovery and muscle training. Meanwhile, further configurations are in development for use in immersive virtual and augmented environments.

“Our goal with the Digits modules is to redefine human-machine interaction through reconfigurable robots that adapt their shape, stiffness, and haptic feedback,” said Paik. “This adaptability supports more tangible virtual reality, effective rehabilitation, and richer experiences for everyone—regardless of their size, ability, and need.”

The study appears in Advanced Intelligent Systems.