According to the World Bank, 98% of the waste people produce in many places, including the United States, ends up in landfills, even though 80% of which is recyclable. As a consequence, more trash is unnecessarily being burned in incinerators and converted into air pollution.
Since it is estimated that global generation of solid waste is going to grow threefold (to 11 million tons per day) by 2100, posterity will likely inhabit a world akin to Wall-E’s fictional Earth.
Most of the planet has been having a serious garbage management problem for the longest time, but many countries are feeling extra pressure to recycle lately after China banned the importation of recyclable waste in 2017.
Considering the urgency of the issue, many tech companies have made it a priority in the past months. One of the innovative solutions that made the news is RoCycle, a robot designed to sort raw garbage automatically.
Grab, Squeeze, Throw, Repeat
Developed by the researchers from Massachusetts Institute of Technology and Yale University, this bot is equipped with Teflon arms to conform to the shape of the trash for easy grabbing, holding, and squeezing and with sensors to gather data in order to recognize an object’s makeup based on its size and stiffness.
RoCycle’s tactile sensors are also conductive, which helps reveal an object’s material depending on how it affects the electrical signal.
This technology is still in the testing phase and currently shows an accuracy rate of 85% when detecting stationary pieces of garbage. However, its performance drops to 63% when handling objects moving in a conveyor belt.
Furthermore, the researchers said that much of the inaccuracy does not necessarily stem from analyzing objects in motion. Rather, RoCycle is still struggling to know where products with different layers of material, like aluminum items covered with paper, should go.
Rocycle is not the first of its kind. In fact, many artificial intelligence–driven bots have been rolled out.
TrashBot is one of them. It is a large container equipped with cameras and sensors to sort out trash automatically in commercial settings. This bot uses machine learning to improve its accuracy in garbage detection and segregation.
Another AI-powered bot is Oscar, which is designed to tell a person whether a piece of trash should go to a recycling facility or a landfill. Its 32-inch display uses computer vision to analyze the items held by a user’s hand and to provide disposal advice.
The folks behind RoCycle are planning to play with more sensors, similar to those utilized by other recycling bots. In future experiments, RoCycle’s sense of touch will be complemented by a vision system to use sight to make up for the machine’s inherent weaknesses.
The initial buyers of the recycling robots on the market are airports and different members of the private sector. With widespread adoption of AI-powered automated machines to improve the efficiency and accuracy of garbage segregation, more players in the recycling space will gain more resources to stimulate their business.
If proven commercially viable, hopefully it will create a virtuous cycle of waste management before the planet’s mess can become too great to sort out.