Bulk Handling Systems (BHS) has launched its artificial intelligence (AI) that identifies recyclables and other items for recovery.
Max-AI technology uses ‘multilayered neural networks’ and a vision system to see and identify objects similar to the way a person does. The technology is intended to drive improvements in MRF design, operational efficiency, recovery, system optimisation, maintenance and more.
The first machine utilising Max-AI technology is an autonomous quality control (QC) unit that sorts container streams following optical sorting. This robotic sorter uses its vision system to see the material, its AI to think about and identify each item, and a robot to pick targeted items.
It is designed to make multiple sorting decisions – for example, separating materials such as thermoform trays, aluminium and fibre while removing residue from a stream of PET bottles. All this is done at rates exceeding human capabilities.
According to BHS chief executive Steve Miller: “Labour is a significant challenge for MRF operators and it is obvious that Max will be beneficial in helping our customers to manage that aspect of their business.
“But the highest returns will come from complete integration of Max-AI technology throughout every advanced BHS system.
“Our customers will not only have autonomous sorting, but also an intelligent central nervous system that observes what is happening in the plant in real-time and adjusts process parameters to maximise profits.”
The first commercial QC unit is already in operation at Athens Services’ MRF in Sun Valley, California. Athens was an ideal location as the Max-AI robotic sorters complemented the advanced screen, air and optical separation technology already in use. Integrating with the company’s existing near infrared optical sorters, Max provides a fully autonomous PET sorting solution.
“This technology was simply not possible until now,” said Thomas Brooks, BHS director of technology development. “Recent advances in computer processing capabilities have enabled us to develop this machine-learning platform.”
Max is central to BHS’s plan to bring autonomous optimisation to MRFs, increasing their performance and profitability. Roy Miller, vice-president of engineering, considers this to be revolutionary for the recycling industry.
“This is the culmination of decades of technological development in recycling,” he said. “Operating costs will go down while uptime, throughput, recovery and purity will all increase, leading to significant economic benefits for our customers and environmental gains for stakeholders.”
Max is able to identify materials in a MRF by processing video images through a detection pipeline and its learning neural networks. These networks are computational machine-learning models based on distributed representation that are inspired by the architecture of the human brain.
“This technology was simply not possible until now. Advances in computer processing capabilities have enabled us to develop this machine-learning platform.”
Such systems excel in areas where the solution or feature detection is difficult to express in a traditional computer program. Max uses self-learned behaviour, making a decision based on the neural network’s inference.
The neural network solves problems much in the same way that the human brain does where, from birth, humans process many images with their eyes, learning to identify objects.
Only possible with the latest AI algorithms and high-performance computing, Max is trained with millions of images that have different materials already identified. The system discovers the best path to take through its artificial brain to reach the correct answer.
Through millions of iterations, Max learns how to identify new images, correctly classifying objects never seen before.
While the first equipment powered by Max-AI technology is the autonomous QC in a container role, BHS envisions a system, rather than simply a robotic sorter, benefiting from AI. As other tools are developed and empowered with the technology, the company plans to roll out autonomous machines in QC roles such as fibre and pre-sort functions.
Max-AI technology will eventually do things even the best plant operators can only dream of accomplishing. As development continues, the technology can also be used for market information, commodity prices and purity requirements to make the ideal bale. BHS believes the possibilities of the technology are game-changing.
Peter Raschio is marketing manager at Bulk Handling Systems