Researchers have developed a groundbreaking sensorized robotic hand that integrates high-density tactile sensing, dexterous soft fingers, and AI-driven feedback for advanced manipulation and human-robot interaction.
Study: Soft robotic hand with tactile palm-finger coordination. Image Credit: aerogondo2/Shutterstock.com
Addressing Limitations in Robotic Dexterity
For years, robotic hands have struggled with dexterity and adaptability. Traditional designs relied on rigid structures with distributed sensors, making them complex and limited in their ability to interact naturally with objects.
Soft robotics offered a more flexible approach, but early designs mainly focused on finger-based sensing, missing out on the full potential of tactile feedback. Some integrated tactile sensing into the palm, but the density wasn’t high enough for fine-grained interactions. The goal? A robotic hand that combines soft, dexterous fingers with advanced tactile sensing to handle objects more like a human hand.
Published in Nature Communications, this breakthrough robotic hand features a high-density tactile palm and flexible fingers that work together seamlessly. The tactile palm enhances contact sensing through a visual-tactile system, while fiber-reinforced pneumatic fingers enable a wide range of grasping techniques.
This powerful combination improves grasp stability, surface reconstruction, and object classification, making the hand ideal for tasks like flaw detection and dynamic pose adjustment. The secret ingredient was AI-driven feedback that allows real-time adjustments, creating a more responsive and adaptable robotic system.
How It’s Built: The Tech Behind the Touch
Soft Robotic Fingers
The robotic fingers were crafted using a high-precision 3D printer (Bambu Lab X1) to create molds and endcaps from polylactic acid. The inner structure was cast with Dragon Skin 10 silicone rubber, ensuring both durability and flexibility. A reinforcing thread was wrapped around a polyethylene fabric mesh at the base to maintain structural integrity. The outer layer, made from Ecoflex 00-30 silicone rubber, gives the fingers a soft, adaptive surface for a secure grip.
Tactile Palm Structure
Unlike most robotic hands that rely solely on finger sensors, this design features a highly sensitive palm. Rigid polylactic acid components house a compact camera and LED ring, while a sensing elastomer made from Ecoflex 00-30 is bonded with an acrylic plate to ensure optical clarity. A layer of gray silicone ink boosts contrast in tactile images, and a final coating of silicone oil enhances surface sensitivity. A diffuser panel, secured with silicone gel, makes sure the palm remains durable—even in high-contact scenarios.
Smarter Control for Better Performance
A pneumatic control system fine-tunes finger movement using proportional regulators, allowing for precise pressure adjustments. Real-time control is handled through a dSPACE DS1103 board and a Matlab-based interface, ensuring the right amount of force is applied for every grasp. Tactile data is processed through image analysis, converting pressure distributions into 3D surface reconstructions using a Fast Poisson Reconstruction algorithm.
In real-world tests, the robotic hand showcased its ability to grasp various objects, classify different fabrics, and detect surface defects. By analyzing color variations in the hue, saturation, and value (HSV) spectrum, the system could dynamically adjust its grip. A ResNet34 model trained on collected tactile images allowed the hand to recognize materials with high accuracy, while a color difference metric pinpointed fabric flaws. These features give the hand a level of perception that makes it highly effective in unpredictable environments.
TacPalm SoftHand: Bringing Robotic Touch to Life
The TacPalm SoftHand is designed to feel and function like a human hand. It combines a high-density tactile palm with soft fingers that can power-wrap around objects or execute precision pinches. Its visual-tactile system uses a micro camera and multi-layer sensing body to detect fine contact details. A built-in LED ring, programmed with red, green, and blue lights, provides depth information, allowing for more accurate object interaction.
The soft fingers, made from fiber-reinforced elastomeric chambers, feature two independently actuated segments, mimicking natural finger movement. A custom-designed palm-finger connector, with three symmetric bases set at a 40° inclination, ensures a smooth, coordinated interaction between the palm and fingers. The end result is a robotic hand that moves and feels more like a human one.
Putting It to the Test
To evaluate its sensing performance, researchers pressed standard probes against the palm’s surface using a motion platform equipped with a force sensor. The TacPalm SoftHand achieved an impressive sensing density of 181,000 units/cm2—far surpassing the human hand’s 240 units/cm2. It detected pressures as low as 1 kPa and remained stable over 5000 loading-unloading cycles. By analyzing light intensity variations, the palm reconstructed 3D surface shapes with 92.3 % accuracy in the x–y plane and 81 % in depth estimation.
TacPalm SoftHand’s ability to coordinate palm-finger interactions enables it to continuously detect objects and identify defects. During a teapot pouring test, the system used real-time tactile feedback to stabilize its grip and adjust for shifting weight. It also leveraged machine learning to recognize objects in sealed bags with 88 % accuracy, proving its potential for handling complex, unstructured tasks.
Conclusion
The TacPalm SoftHand isn’t just another robotic hand—it’s a game-changer in tactile robotics. With high-density tactile sensing and dexterous soft fingers, it offers precise, adaptable manipulation like never before. The visual-tactile palm captures detailed contact data, while fiber-reinforced pneumatic fingers enable versatile and secure grasps.
By seamlessly coordinating palm and finger movements, this system improves stability, surface detection, and object classification. Add AI-driven feedback into the mix, and you have a robotic hand that’s not just responsive—it’s smart, making real-time adjustments for smoother, more natural interactions.
Journal Reference
Zhang, N., et al. (2025). Soft robotic hand with tactile palm-finger coordination. Nature Communications, 16:1, 1-14. DOI:10.1038/s41467-025-57741-6, https://www.nature.com/articles/s41467-025-57741-6
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.