Combining an innovative pneumatic design, vision-based sensing, and deep learning, the proposed system could find applications in delicate object handling

Soft robotic grippers can safely handle fragile objects, making them promising for applications such as food handling, nursing, and household services. However, giving soft grippers a sense of touch remains a major engineering challenge. Now, researchers from Japan have developed EleTac, an elephant trunk-inspired soft gripper that combines gentle grasping with tactile sensing using a single internal camera. The system can accurately estimate contact location, force applied, object shape, and the position of its fingers.

Soft grippers, which are built from flexible materials that can bend and deform, are attracting a lot of attention from robotics researchers worldwide. Unlike conventional robots made from rigid metal or plastic, soft grippers can grasp items more gently while naturally adapting to different shapes. This makes them uniquely suitable for delicate tasks such as handling fruit, baked goods, lab samples, and medical supplies. 

However, grasping objects is only part of what robots need to do. To manipulate objects as skillfully as humans, robots also need a sense of touch; they need feedback on where contact is happening, how much force is being applied, what shape the object has, and even the position of the gripper's own fingers. The main challenge lies in the fact that the same flexibility that makes soft grippers useful also makes it difficult to equip them with sensors. Because soft gripper materials constantly bend and stretch, conventional sensors are hard to integrate without limiting the gripper's flexibility or only covering small areas. But what if we took a page from one of nature's most impressive grippers?

In a recent study published in the IEEE Transactions on Robotics journal on June 26, 2026, a research team led by Professor Van Anh Ho from the Japan Advanced Institute of Science and Technology (JAIST), Japan, developed EleTac, a soft robotic gripper inspired by the tip of the elephant's trunk. Their work describes an innovative design that combines object grasping with high-resolution tactile sensing and proprioception, which is the ability of a robot to sense the position and movement of its own body. Other members of the team included Associate Professor Shan Luo and doctoral student Xuyang Zhang from King's College London, U.K.; Dr. Tuan Tai Nguyen from JAIST; and Dr. Quan Khanh Luu from Purdue University, USA.

EleTac uses a pair of soft, pneumatically actuated fingers that gently close around objects. Instead of embedding numerous sensors throughout the gripper, the researchers equipped it with a single internal fisheye camera. This camera observes, from the inside, how the soft material deforms during grasping. The team trained several deep learning algorithms on the resulting images to extract many kinds of tactile information. 

The team evaluated EleTac through a series of grasping and perception experiments. Using a simple control strategy, the gripper successfully handled a wide variety of objects, including fruit, tofu, fabric, tools, bolts, and even playing cards. They also tested EleTac in two particularly challenging scenarios: searching for and pulling out a pen buried in sand using touch alone, and wiping marker ink off curved tableware with a sponge. This second task required the gripper to adjust its motion based on how firmly it sensed the sponge pressing against each surface.

The researchers see EleTac's compact, lightweight, and low-cost design as one of its strongest practical advantages. "EleTac, fully covered with vision-based tactile skin, is suitable as a plug-and-play end-effector for existing robotic platforms and humanoid robots. For instance, the gripper shows promise for delicate tasks," remarks Prof. Ho.

Beyond near-term applications in business and industrial settings, EleTac also represents a step towards the widespread adoption of service robots. "With the ability to detect contact and apply gentle forces, soft robots may be able to assist with everyday tasks, support elderly or vulnerable individuals, and perform operations that are difficult for today's rigid robotic systems. We envision safer and more capable robots that can work alongside people in homes, hospitals, and public spaces in the near future," concludes Prof. Ho.

Image title: Concept of the EleTac gripper
Image caption: (a) Inspiration from the elephant trunk. (b) The soft pneumatic gripper closes using air pressure, enabling gentle and versatile grasping. (c) Through vision-based sensing, the gripper can perceive both contact with external objects and its own shape.
Image credit: Professor Van Anh Ho from Japan Advanced Institute of Science and Technology, Japan
License type: Original content
Usage restrictions: Cannot be reused without permission

Image title: Applications of the EleTac
Image caption: The soft gripper can safely handle fragile objects. Its sensing capabilities support tasks such as automated dishwashing and exploration in environments with limited visual sensing.
Image credit: Professor Van Anh Ho from Japan Advanced Institute of Science and Technology, Japan
License type: Original content
Usage restrictions: Cannot be reused without permission

Reference

Additional information for EurekAlert

Funding information

This work was supported by the JST CREST, Japan, Grant Number JPMJCR2554, and the JSPS Grant-in-Aid for Scientific Research (KAKENHI) under Grant 25KJ1365, and the UKRI EPSRC Open Fellowship "TacDiff: Designing Tactile-based Robots via Differentiable Simulations" (UKRI3438).

July 7, 2026