EleTac: A Soft Robotic Gripper Inspired by an Elephant's Trunk Tip, Endowing Robots with Precise Tactile Sensation

The Intersection of Bionics and Soft Robotics

Soft robotics is one of the most dynamic branches of the robotics field in recent years. Unlike rigid metal components, soft robots are made of flexible materials that can naturally adapt to object shapes and grasp fragile items in a gentler manner. However, this flexibility also brings challenges for sensor integration—traditional rigid sensors struggle to follow the large deformations of the material and have limited coverage.

Professor Van Anh Ho’s team at the Japan Advanced Institute of Science and Technology (JAIST) found a solution inspired by one of the most dexterous "grasping organs" on Earth—the tip of an elephant’s trunk. In a study published in *IEEE Transactions on Robotics*, they introduced EleTac: a soft gripper inspired by the elephant trunk tip that uses a single fisheye camera and deep learning algorithms to extract tactile information and proprioception from deformed images.

Core Design Philosophy: Vision as Touch

EleTac consists of a pair of pneumatic soft fingers, with only one miniature fisheye camera placed inside the internal cavity. When the fingers grasp an object, the flexible material deforms, and the camera records changes in the inner surface pattern in real time. The research team trained multiple deep neural networks to decode contact position, grasping force, object shape, and the bending state of the fingers themselves from these images.

This "visual-tactile" strategy cleverly bypasses the challenge of embedding sensors in flexible materials. Traditional tactile sensors often require array wiring, which is both expensive and limits deformation. EleTac, on the other hand, leverages mature camera and algorithm technology to turn tactile perception into a visual computing problem, significantly reducing system complexity and cost.

Performance Verification: From Tofu to Finding a Pen in Sand

In experiments, EleTac demonstrated remarkable adaptability and sensing accuracy. It successfully grasped tofu, fruit, fabric, tools, bolts, and even playing cards, without requiring complex control strategies. The research team also designed two more challenging tasks:

• Finding a pen in sand: Relying solely on tactile feedback, the gripper searched through sand and picked up a buried pen.
• Wiping curved tableware: The gripper held a sponge and, based on perceived contact force, automatically adjusted its motion trajectory to erase ink marks from a curved table surface.

These two tasks simulated uncertain environments that robots might encounter in the real world. EleTac’s performance demonstrated the effectiveness of its tactile feedback in dynamic control.

Industrial Significance of Japanese Soft Robotics Research

EleTac’s outstanding feature lies in its compact, lightweight, and low-cost design. Professor Ho noted: "EleTac can serve as a plug-and-play end effector, suitable for existing robot platforms and humanoid robots." This suggests that soft robots are moving from the laboratory to practical deployment.In Japan, the robotics industry has long held a leading global position, but its focus has mostly been on industrial robotic arms and automated production lines. With the intensification of labor shortages and an aging society, the demand for service robots has surged. The "flexible + tactile" technology route represented by EleTac precisely fills the gap in fine manipulation tasks—such as food processing, laboratory sample handling, medical assistance, and home care.

Furthermore, this research combines computer vision and deep learning, reflecting the vitality of interdisciplinary innovation at Japanese universities. As a national university specializing in graduate education, JAIST has played an important role in such cutting-edge fundamental research, and its collaboration with international teams such as King's College London and Purdue University demonstrates the maturity of Japan's open research ecosystem.

Outlook: From Grasping to Autonomous Behavior

Currently, most robots still rely on pre-programmed instructions or external visual positioning. EleTac's tactile proprioception enables robots to perceive their own interaction with objects in real time, which is a key step toward truly autonomous operation. In the future, combined with stronger AI models, soft robots are expected to make independent decisions in unknown environments—for example, helping the elderly put on or take off clothes, or assisting doctors in surgery.

Professor Ho concluded: "We envision that in the near future, safer and more reliable robots will coexist with humans in homes, hospitals, and public spaces." From the tip of an elephant's trunk to the future of human-robot collaboration, EleTac provides a reusable perception paradigm for soft robots and also showcases Japan's deep expertise in bio-inspired robotics.