Image default
AI

MIT Researchers Create Robotic Gripper to Untangle Thin Cables

MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) has shared the results of a new project in which it built a two-fingered robotic gripper, which has soft pads for dedicated and fine manipulation of objects like cables, sheets and more. The robot’s design is based on how humans use their fingers to do things like untangle wires and tie knots.

To do this, the CSAIL research team equipped their robotic gripper with fingertips that are not only made out of a soft material, but that also have embedded sensors which help it continually detect the position of a cable between the grippers to better control holding and manipulating them while performing simple tasks like detangling.

The fingertip sensors provide high-resolution tactile information, using so-called “GelSight” technology that embeds tiny cameras in soft rubber. These sensors provide data on how the cable is situated between the two ‘fingers’ of the gripper, and on how much force is being exerted on the cable in terms of friction as it moves between the pads. This allows it to change its pose and grip depending on what’s needed to get the cable into the position you want – an approximated version of what we’re doing when we work with cables or cords.

The CSAIL team, led by MIT post-doc student Yu She, was able to demonstrate the gripper following a USB cable along its length from any random starting holding position, and could also work its way down the length of a cable with a ‘hand-over-hand’ motion when working in tandem with a second gripper, in order to find the end of the cable. The system demonstrated its ability to work across different types of cables, including different materials and lengths.

CSAIL’s system managed to perform a task that is very commonplace for us, but that poses a significant challenge or a robot – plugging a set of earbuds into a phone’s stereo headphone jack. It may not seem like much, but it opens the door for building out the technology to perform ever more sophisticated tasks, including doing things like folding cloth, tying knots, and ultimately even potentially sewing sutures to close wounds during medical procedures.

The goal, according to She, is to create a robot that can handle this delicate work in order to provide a safe alternative when having a human do the same would be potentially dangerous, as well as “repetitive” or “dull.” Like many robotics endeavors, you can see how creating a robot that can handle this kind of work could free up human time to focus on more complex or advanced tasks.

The team intends to look at applications in the auto industry first – a good target not only because automation is already a key part of automotive manufacture, but also because wiring and threading cables represents a significant portion of remaining manual work in car production.

Related posts

Lunit Gets FDA Nod for AI-based Chest X-ray Triage Solution

PR Newswire

Juniper Networks announces Apstra Freeform

Business Wire

Tenstorrent Selects SiFive Intelligence X280 for AI Processors

Business Wire