Source: MITA technique that enables robots to sort recyclables by touch has been developed by researchers from Yale University and MIT.

Unlike other current recycling robots now on the market, which rely on computer vision to distinguish among materials, MIT and Yale’s RoCycle system distinguishes among paper, plastic and glass using soft robotics and sensors. The system attaches to any robotic arm and includes a gripping appendage composed of a material called auxetics, which expands when pulled. The material enables the robotic hand to conform to an object’s surface and a sensor determines an object’s size, while pressure sensors gauge the amount of force exerted to grasp an object.

During testing, RoCycle demonstrated 85% accuracy when identifying and sorting the three materials from a fixed position. Accuracy decreased to 63% when RoCycle collected items from a moving conveyor belt.

According to the researchers, an entirely optical, object-sorting process often results in inaccuracies as material type is a tactile property and not always a visual one. For example, many previously disposable items, such as plastic cutlery, are now being manufactured in durable versions that resemble their metal cutlery counterparts, making it difficult to distinguish the material by sight alone.

“Computer vision alone will not be able to solve the problem of giving machines human-like perception, so being able to use tactile input is of vital importance,” MIT professor Daniela Rus said.

To further improve RoCycle’s accuracy, the research team intends to add a camera and computer vision to the RoCycle system.

The research will be presented later this month at the IEEE International Conference on Soft Robotics in Seoul, South Korea.