Researchers from  MIT's Computer Science and Artificial Intelligence Lab (CSAIL) and Yale University have developed a robot, known as RoCycle that can sort through recycling by picking up paper, plastic, and glass. RoCycle uses an arm that's bundled with soft grippers and sensors to identify what each material is made of through touch. The research will be presented later this month at the IEEE International Conference on Soft Robotics in Seoul, South Korea. It's also supported by companies like Amazon, JD, the Toyota Research Institute, and the National Science Foundation.


RoboCycle's arm sorts trash by using soft grippers and sensors that distinguish between paper, metal, plastic and glass. (Image Credit: Jason Dorfman, MIT CSAIL)

Most recycling centers already use magnets to pull out metals and air filters to separate papers from heavy plastics. Even without those methods, most of the sorting is done hand, which can be dirty and dangerous for the worker. Robotic grippers are an improved method for recycling automation, mostly because it's a natural process. If recycling isn't properly sorted, the waste ends up in landfills, where 25% of all waste is contaminated.


Using computer vision on the robot will not be enough for the system to sort through the recycling objects like a human would, so using tactile sensors would give it that perception. The computer vision method is also inaccurate since material type isn't a visual property, but rather, it's tactile.


The soft grippers on RoboCycle are a pair of cylinders and high-deformation capacitive pressure and strain sensors. Researchers plan on installing a camera and computer vision to RoboCycle's tactile sensors in an effort to improve its accuracy. The system was only 85% accurate in identifying and sorting the three materials by touch when in a set position. Tests were conducted with 27 objects using the robot, and it was only 63% accurate when materials were gathered from a conveyer belt.


The RoCycle system can also be installed on any robotic arm. Its gripping appendage is made out of customizable material called auxetics that widens as it's being pulled on. This material also enables a robotic hand to fit onto an object's surface, twisting it around as it's being cut. A sensor on the gripper identifies the object's size while a pressure senor measures how much force is being used to take hold of the object. Since the sensors are conductive, it can also detect metal.  Data collected from these sensors are used to determine the type of material the arm can pick up.


Proper disposal and sorting of recycling materials have reached a high level of importance since last year when China announced they would no longer be receiving plastic recycling imports in 2017.


These types of robots could be used in apartment buildings or university campuses to carry out sorting tasks of people's recycling.



See the video here:


Have a story tip? Message me at: cabe(at)element14(dot)com