RobotSweater, developed by a research group from Carnegie Mellon University’s Robotics Institute, is a machine-knitted textile “skin” that has the potential to sense pressure and contact.
RobotSweater, developed by a research team in the Robotics Institute and shown here on a robotic arm, is a machine-knitted textile “skin” that can sense contact and pressure. Image Credit: Carnegie Mellon University
Similar to knitters that can take any type of yarn and turn it into a hat, sock, or sweater of any size or shape, it is possible to tailor the knitted RobotSweater fabric to suit ununiform three-dimensional surfaces.
“We can use that to make the robot smarter during its interaction with humans,” stated Changliu Liu, an Assistant Professor of Robotics at the School of Computer Science.
Knitting machines can pattern yarn into shapes that are non-flat, that can be curved or lumpy. That made us think maybe we could make sensors that fit over curved or lumpy robots.
James McCann, SCS Assistant Professor, Carnegie Mellon University
In the past few years, McCann’s study has concentrated on textile fabrication.
As soon as it is knitted, the fabric could be utilized to help the robot “feel” when a human touches it, especially in an industrial setting where safety is paramount. Present solutions available for detecting human-robot interaction in the industry look like shields and utilize highly stiff materials, which Liu notes cannot cover the robot’s complete body since parts need to deform.
“With RobotSweater, the robot’s whole body can be covered, so it can detect any possible collisions,” stated Liu, whose research focuses on industrial applications of robotics.
The knitted fabric of RobotSweater comprises two layers of conductive yarn made with metallic fibers to conduct electricity. Fixed between the two is a net-like, lace-patterned layer. On being subjected to pressure on the fabric—say, from someone touching it—the conductive yarn closes a circuit and it is read by the sensors.
The force pushes together the rows and columns to close the connection. If there's a force through the conductive stripes, the layers would contact each other through the holes.
Wenzhen Yuan, SCS Assistant Professor and Director, RoboTouch lab, Carnegie Mellon University
As well as how to design the knitted layers, requiring dozens if not hundreds of samples and tests, the team faced one more difficulty in linking the wiring and electronic components to the soft textile.
There was a lot of fiddly physical prototyping and adjustment. The students working on this managed to go from something that seemed promising to something that actually worked.
James McCann, SCS Assistant Professor, Carnegie Mellon University
What worked, in the end, was wrapping the wires around snaps fixed to the ends of every stripe in the knitted fabric. Snaps are an economical and efficient solution, such that even hobbyists making textiles with electronic elements, called e-textiles, could utilize them, stated McCann.
“You need a way of attaching these things together that is strong, so it can deal with stretching, but isn't going to destroy the yarn,” he said, adding that the team also discussed using flexible circuit boards.
After fitting into the robot’s body, RobotSweater will develop the potential to sense the distribution, shape, and force of the contact. Also, it is highly precise and efficient compared to the visual sensors that the majority of robots depend on at present.
Yuan stated, “The robot will move in the way that the human pushes it, or can respond to human social gestures.”
In their study, the team illustrated that pushing on a companion robot outfitted in RobotSweater told it which way to move or what direction to turn its head. When utilized on a robot arm, RobotSweater enables a push from a person’s hand to guide the movement of the arm, while grabbing the arm told it to open or close its gripper.
In research that will be performed in the future, the team wishes to explore how to program reactions from the swipe or pinching motions utilized on a touchscreen.
The team—such as SCS Ph.D. students Zilin Si and Tianhong Catherine Yu, and visiting undergraduate student Katrene Morozov from the University of California, Santa Barbara—will present the RobotSweater research paper next week at the 2023 IEEE International Conference on Robotics and Automation (ICRA).
Begun by the three faculty members in a conversation over lunch one day, the partnership and specialties among the team of scientists helped the RobotSweater come to life, stated McCann.
McCann stated, “We had a person thinking about fabrication, a person thinking about the robotics integration, a person thinking about sensing, and a person thinking about planning and control. It's really nice to have this project where we have the full stack of people to cover each concern.”