AI-powered robots used to pick valuable materials from Fife waste

Zoe Cook, of Recycleye and David Goodenough, of Cireco, with one of the robots at Cireco

AI-powered robots are being used to sort Fife Council’s household waste in its Dunfermline material recovery facility (MRF).

Installed in late 2022, the robotic pickers extract valuable recyclables from the residual line at the end of the plant, with the material otherwise being sent to landfill. Resource management service, Cireco Scotland, is utilising the robots on behalf of the local authority, with the technology developed by London-based Recycleye.

Fife Council’s MRF sorts 30,000 tonnes of household recyclables annually, such as aluminium, plastics, paper and cartons. The pure material streams are then compacted into bales and sent for chemical or mechanical reprocessing into recycled material.

Previously, the materials had been sorted by analogue machinery such as magnets, near-infrared sensor sorters, and a team of manual pickers who perform quality control at the end of the process, picking out any recyclables that have been missed.

However, prior to September 2022, Cireco said that it noted that the ‘residual’ line at the end of the plant, which should comprise non-recyclable material to be sent to landfill, still featured a large amount of target materials – particularly PET bottles and aluminium cans. The firm said this is ‘highly valuable’ materials, which can be resold by Cireco at high costs when packed into pure material bales, supporting the local economy.

The high value led to the resource company investing in three Recycleye robots to automate and increase recovery of target recyclables, capitalising on their value rather than losing them to landfill.

The AI-powered robots are designed and manufactured by Recycleye, a technology company delivering AI-powered sorting equipment to waste management businesses globally, which has installed the waste-sorting robots 30 times across the UK and wider Europe.

The technology comprises a computer vision system which sits atop the facility’s conveyor belt, leveraging a camera and machine learning algorithms to detect each waste item by material and object e.g. aluminium beverage can. This detection information, along with the coordinates of which bin the item should be sorted into, is sent to the corresponding robotic arm, which physically picks and shoots the material accordingly.

As a result of this innovation, Cireco has relocated the three manual pickers from this line (six FTEs across two shifts) to engineering maintenance roles and other lines within the plant, meaning increased value is being derived from other material streams too. Meanwhile, the robots pick 30,000 items consistently across a shift, so subsequently there is more regular recovery of target materials from the residual line.

David Goodenough, service manager corporate operations and projects at Cireco, said,Recycleye is very forward-thinking and worked with us to develop a system that would work for our unique set of materials that we come across and the batch process that we operate. We are proud that this innovation is supporting increased recycling rates for Scotland, and ensuring resources are fed back into the local economy.”

AI also has the ability to be flexible. The composition of waste received at the Dunfermline MRF varies depending on season, trend and packaging changes. When the robots were initially installed, there was a small number of aluminium cans reaching the final QC stage, compared to a higher proportion of PET bottles. Hence, only one robot was programmed to pick both aluminium and PET, with the remaining two only picking PET bottles. Yet, in subsequent months, increasingly more aluminium cans were reaching the line, hence Recycleye reprogrammed the system so that all three robots were picking both PET and aluminium.

Equipping Cireco with flexible equipment will be important as the Scottish and UK governments introduce new recycling regulations, Cireco said. The data that Cireco receives from the AI systems on material output composition will also contribute to compliance reporting for these new regulations.

Zoe Cook, technical sales manager (UK) at Recycleye, added, “This robotics installation at Cireco is a prime example of the flexibility of AI to adapt to compositional and regulatory changes, and Scotland’s waste management businesses that are leading the way in innovation.”