Robots may be coming for your chores someday, but at the University of Texas at Austin they are starting with something far trickier: picking up a potato chip without turning it into crumbs.
Researchers at UT Austin have built soft robotic hands that can grab extremely delicate items, from eyeglasses and raspberries to those fragile chips, without crushing them. The system, called FORTE (Fragile Object Grasping with Tactile Sensing), combines 3D-printed, fin-ray-inspired fingers with embedded air-channel tactile sensors so the robot can actually feel pressure and slip. In lab tests, the grippers handled tasks that typically stump robot hands that rely only on vision.
According to UT Austin News, the work appears in IEEE Robotics and Automation Letters and is led by doctoral student Siqi Shang, with collaborators Yuke Zhu, Mingyo Seo and Lillian Chin. As reported by KXAN, the team put the grippers through trials with 31 different objects and logged a roughly 91.9% single-trial success rate, outperforming grippers that only use cameras. The researchers say the system's fast slip detection is the critical upgrade that lets it handle fragile items instead of pulverizing them.
How FORTE senses touch
FORTE's fingers are built with internal, empty air channels that deform when the robot makes contact with an object. Off-the-shelf sensors read those tiny pressure changes and feed the information back to the controller, which lets the system estimate grasp forces and detect slip in real time.
The paper on arXiv reports that the team can estimate forces from 0 to 8 newtons with an average error of about 0.2 newtons and can detect slips in roughly 100 milliseconds. That tactile feedback loop lets the gripper add only the minimum extra force needed to stop an object from slipping, instead of simply squeezing harder and risking damage.
Lab tests and performance
To see how FORTE handled real-world challenges, the team tested it in single-trial experiments with 31 fragile, slippery and everyday items. Jam jars, billiard balls, raspberries and potato chips all went into the mix. The system achieved success rates near 92% and reported high slip-detection accuracy, according to Axios.
The researchers say that precise slip sensing is key to avoiding false alarms that would otherwise cause a robot to clamp down harder and crush a delicate object. Axios also notes that the team has publicly released the hardware designs and algorithms so other researchers and developers can build on the work.
Why it matters in Austin and beyond
UT Austin researchers see immediate applications for FORTE in food processing, manufacturing and health care, especially for tasks that require a very light touch. The project received support from the Texas Robotics Industrial Affiliate Program, the National Science Foundation, the Office of Naval Research and DARPA's TIAMAT program, according to UT Austin News.
The team is now working to fine-tune the sensors so they are less sensitive to temperature changes and to improve the gripper's ability to catch objects as they start to slip. By open-sourcing the hardware and code, the researchers hope that others will adapt FORTE for local startups and manufacturing partners in Central Texas.
FORTE is not about to show up in the grocery aisle stocking shelves tomorrow, but it marks a practical, incremental step toward robot hands that can do household chores and assist people without breaking whatever they touch. For Austin, it is another case of lab research that could spin out into commercial projects and startups focused on tactile robotics.
