ARTiS: An Adaptive Robotic Gripper for Enhanced Tool Manipulation in Disassembly Applications
This paper introduces ARTiS, a novel adaptive robotic gripper that combines soft mechanisms, anthropomorphic dexterity, and rigid robustness to securely manipulate various tools in disassembly applications through active jamming and fin-ray adaptation.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Robots have become remarkably good at picking things up. In factories, they can lift boxes, sort fruit, and move heavy parts with a precision that often surpasses human speed. However, there is a distinct difference between simply holding an object and actually using it. When a human picks up a hammer, a screwdriver, or a power drill, they do not just clamp down on the handle; they wrap their hand around it, adjust their grip as the tool moves, and use their palm to lock the tool in place so it does not spin or slip while force is applied. This ability to hold a tool securely while twisting, pushing, or striking is a fundamental part of how we build and take things apart. For robots, this remains a stubborn hurdle. Most robotic hands are designed to grab objects, but when they try to use a tool, the tool often slips, spins, or falls because the robot lacks the complex, adaptive grip that a human hand provides naturally.
A team of researchers has developed a new robotic hand designed specifically to solve this problem. They call it ARTiS, a three-fingered gripper that mimics the way a human hand holds and uses tools during disassembly tasks. The core idea behind the design is to combine two different ways of holding things. The first part is a soft, jamming palm. Imagine a bag filled with tiny glass beads; when air is sucked out of the bag, the beads pack tightly together, turning the soft bag into a rigid shape that conforms perfectly to whatever object is inside. The researchers built a palm like this, but shaped it with a deep slot so that tools like screwdrivers or hammers can be pushed deep into it and locked in place. The second part consists of three fingers that can move independently and are tipped with special, flexible pads. These fingertips are designed to bend and adapt to the shape of the tool handle, providing extra grip and stability. By using the jamming palm to hold the main body of the tool and the fingers to adjust the tip, the robot can hold a tool securely enough to twist a screw or strike a surface without the tool slipping out of its grasp.
The researchers tested this new gripper by asking it to perform the kinds of tasks a human would do when taking apart a machine, such as an automotive air conditioning unit. They used a variety of common tools, including nine different types of screwdrivers, two hammers, and two power drills. The tests involved not just picking the tools up, but also using them to unscrew bolts, break plastic connections, and drive nails. The results showed that the combination of the jamming palm and the adaptive fingers worked significantly better than previous robotic hands. When the researchers measured how much force could be applied to the tip of a screwdriver before it slipped, the new gripper could handle up to 7.8 times more force than a standard robotic hand that relied only on fingers. In many cases, the palm alone provided over 90 percent of the holding power, proving that the "jamming" mechanism was the key to keeping the tool steady.
However, the study also revealed where the technology still faces challenges. While the gripper was excellent at holding stationary tools or turning screws, it struggled slightly with tools that vibrated heavily, such as a power drill. The vibrations from the drill caused the tiny beads inside the jamming palm to shift, which sometimes loosened the grip and caused the tool to lose contact. Despite this, the gripper succeeded in grasping and reorienting tools in 85 percent of the attempts, but the success rate for actually using the tools was 68 percent. In a collaborative experiment, the gripper held a workpiece steady, allowing a human to use both hands to hold the tool and the screw simultaneously; this setup reduced the total time needed to unscrew three bolts by nearly 14 percent compared to a human working alone. The researchers also tested the gripper on a wide range of everyday objects, from food items to kitchen utensils, and it successfully grasped and lifted 95 percent of them, showing that its design is not limited to just tools.
The project also explored how humans and robots could work together more closely. In one experiment, a person used a "teaching mode" to physically guide the robot's hand into the correct position to hold a part of a machine. Once the robot learned the position, it locked the part in place, allowing the human to use both hands to work on the machine. This collaboration reduced the total time needed to unscrew three bolts by nearly 14 percent compared to a human working alone. The study concludes that while the gripper is not perfect, particularly with vibrating tools, it represents a significant step forward. It demonstrates that for robots to truly master the art of taking things apart, they need more than just strong fingers; they need a smart, adaptable palm that can lock tools in place, just as a human hand does.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.