Robot recycling sorters
We toss our recycling into a single container. Then a truck hauls it to a materials recovery facility (MRF).
Sorting it is monotonous and potentially dangerous work for humans.
Some manufacturers have developed robotic sorting systems.
This infant technology has enormous potential. Robots don’t slow down. They don’t need breaks or vacations (except occasional down-time for routine maintenance). They can work continuously all day and night.
With artificial intelligence (AI), they can learn to separate, say, one plastic from another, as well as human sorters. So far, though, it doesn’t appear that robotic sorting can completely eliminate the need for human eyes.
AI promises to reduce a MRF’s operating costs, speed up their processes, and improve the quality of their output. Can it work economically?
Three robotic recycling manufacturers
ZenRobotics
Robotic waste sorting started in 2011. ZenRobotics, a Finnish company, combined AI, computer vision, and machine learning to operate robotic arms. It built recycling stations and connected them in a neural network.
Real-time feeds provided data from various sensors, including light sensors, metal sensors, and a high-resolution camera. Using AI, the robots learned to select and sort specific items.
ZenRobotics equipment can sort wood from concrete at more than 60 picks per minute. It can learn to recognize new materials from examining 200 samples.
At present, its system mostly sorts materials for construction, industry, and other commercial applications. This waste’s resale value exceeds that of residential recycling.
Sadako
Sadako, a Spanish company, recently developed a similar system. It’s called Max-AI.
When the first prototype of Max-AI was demonstrated, it could only pick PET off a conveyor belt. And it could perform only 20 picks per minute. The latest version separates a much larger range of materials at 80 picks per minute. It finally exceeds human capabilities.
Bulk Handling Systems (BHS) has formed a joint venture with Sadako and Athens Services in Sun Valley, California. Max-AI now operates at Athens’ MRF.
It both sorts various materials and removes residues from PET bottles. BHS anticipates that Max-AI technology will ultimately control the other equipment in its systems.
AMP Robotics
“Clarke” by AMP Robotics
AMP Robotics has begun a pilot project in a MRF in Denver.
It a model of robot that other industries have used for almost 20 years.
The robot’s nickname is “Clarke,” after the science fiction writer Arthur C. Clarke.
The company first trained Clarke to remove cartons from a conveyor belt. It can transfer everything Clarke has learned to other robots at other MRFs.
Now it’s training Clarke to sort plastic. At first, it will concentrate on PET and both natural and colored HDPE (numbers 1 and 2 in the recycling triangle).
Eventually the company hopes to use Clarke to sort plastics coded 3-7. It even hopes Clarke can identify now unrecyclable films and single-serve coffee pods.
Clarke can already remove 60 PET items per minute. Humans can pick 80 items per minute for a short time. But then they get tired and slow down. The average human rate is 40 picks per minute.
The economics of the current recycling system
Commercial recycling has a higher resale value than residential recycling.
For one thing, it usually has less contamination.
Before the 1960s, American households only had to separate wet garbage from dry trash.
And then deal with two different companies with two different collection schedules.
When burning paper became illegal in Los Angeles, households there faced a third company to haul the paper.
Sam Yorty became mayor of Los Angeles on a promise to relieve the burden. His plan let residents put all their waste in one container for one collection. The idea quickly spread across the country.
By the time municipal recycling became widespread, Americans had become accustomed to putting everything in one container. We have resisted sorting ever since.
Municipalities have developed various methods of collecting recyclables.
Jurisdictions that require residents to separate glass, metal, plastic, and paper for recycling have a very low participation rate.
Most operate “single stream” recycling. Residents need only separate recyclables from landfill trash. That method results higher participation, but it introduces more contamination.
Food residue (or water) dripping out of cans and bottles renders otherwise valuable office paper unfit for the most profitable uses.
The rules for what to put in recycling and what to put in trash keep changing. They vary somewhat from place to place. Whether from confusion or carelessness, people very often put something in the wrong container.
Montgomery, Alabama tried out a MRF that didn’t require residents to separate anything. The experiment failed badly. Montgomery and other municipalities that used the MRF couldn’t recycle anything for a year after it closed.
Commodity prices plunged, and Montgomery’s MRF could not get enough for its more contaminated output to stay in business.
The easier we make recycling for consumers, the more expensive it becomes to process it and the less money MRFs can make selling it.
A way to make robotic sorting more affordable
Pilot programs at single-stream MRFs demonstrate that robotic sorting with AI works.
As the technology matures, it will become less expensive to install and operate.
But can it be cost-competitive with human sorting?
Costs for residential recycling could come down if we could adopt a different method of separating trash at home.
In some parts of the UK, households separate organic and dry wastes into color-coded bags. Such a system means that municipalities need not use one truck for garbage and another for recycling.
Robots could easily separate the bags. MRFs could set aside organic waste for composting and sort recyclables. That would send minimal material to landfills or incinerators.
In other words, we could make robotic recycling more affordable for municipal recycling. We just need to go back to something like America’s pre-1960s practices.
But with a single container at the curb.
The post Recycling with robotic sorting and artificial intelligence appeared first on Sustaining Our World.
