Did you know that trash companies sift through 68 million tons of recycling waste each year? This weight is equivalent to more than 30 million vehicles. This process is undertaken by workers that sort out trash like glass, plastic, and paper that spread out in conveyor belts. Such jobs can often be unsafe and it can get difficult to sort out materials.
For instance, workers in the plastic recycling industry might find it strenuous to remove plastic from the mix of waste that comes in. Keeping this issue in mind, a team of researchers from MIT has designed and developed a highly functional robotic system known as the RoCycle. It can easily detect if an item is plastic, metal or paper.
This system includes a Teflon hand that can detect an item’s stiffness and size with tactile sensors on its fingertips. This system is designed to be compatible with any robotic arm. During tests, they discovered that it was 85% accurate in detecting materials when they were stationary and 63% accurate on a conveyor belt. What’s even more surprising is that it could distinguish between two Starbuck cups that were visually similar but different in material. One was made of plastic and the other of paper.
Daniela Rus, who is one of the researchers from the team, stated that the robot’s skin has multiple sensors that provide haptic feedback to sort out a wide range of objects, from squishy to rigid. She further added that tactile input is of prime importance. This is true especially since computer vision alone cannot determine the type of material. It must have human-like perception such as touch to understand the nature of the material so that it can determine its type.
This project aims to minimize the back-end cost of recycling and reduce the safety risks that come with manual labor. The current recycling machines are not completely automated. They use optical sorters, magnetic sorters, and aluminum sorters. The optical sorters make use of wavelength light to differentiate between materials. The magnetic sorters can separate iron and steel products whereas the aluminum sorters use eddy currents to take out non-magnetic items. This issue of having multiple sorters can be resolved if RoCycle were to be deployed on a wide scale across many recycling industries.
With its built-in sensors, Rocycle can determine the radius of material with 30% accuracy and know the difference between soft and hard materials with 78% accuracy. The robotic arm was also designed to be puncture-resistant. The team is now looking to further enhance Rocycle’s functionality by improving its accuracy in determining materials. They plan to do this by combining tactile data with video data by using the robot’s camera.
If you are in the plastic recycling industry or the robotics industry, then Plastivision is a great platform for you to showcase your product range and network with industry professionals. Book your booth today!