AI can help clean up our landfills, says expert

In this week’s Net Hero Podcast, we spoke to Carling Spelhaug of AMP Robotics who tells us about the value of AI and robotics in recycling

Featured Video Play Icon

Robots can help do the jobs humans should never have to do.

This is what Carling Spelhaug, Corporate Communications Director at AMP Robotics told us in this week’s Net Hero Podcast.

AMP Robotics uses AI and advanced technology to sift through recycling to ensure that recyclable materials don’t end up in landfills.

Carling said: ‘At AMP Robotics, we are using computer vision and deep learning to automate the sorting process of mixed material streams.

‘Typically we put everything in a bin and it’s sent to what’s called the materials recovery facility and everything there needs to be sorted into separate commodities so that it can be sent to end markets where people want to buy it to make it into new materials.

‘And the way that this industry has historically run has been on manual labour. So it’s a dull, dirty, dangerous task and one that is ripe for automation.’

Carling told us how dire the rate of recycling is globally, this being especially true for America.

She said: ‘It’s a pretty dismal recycling rate of just about 32% [in America], according to the EPA. We do have the ambitious goal of reaching 50% by 2030.’

She added: ‘There are a lot of factors but I think access is an issue.

‘Even in cities, it can be difficult. But if you think of rural areas in America where if there is recycling collection, it has to be trucked back to the city or somewhere that has a recovery facility where materials can be sorted. And the economics just don’t pencil out on that.’

AMP Robotics’ system helps cut down on the amount of recyclables that end up in landfills or incinerated.

Carling said: ‘We’ve built up this database, it’s the largest known database of recyclable materials for machine learning. And every AMP installation has access to this database.

‘It’s a camera that is a part of a modular system that sits above a conveyor belt and it takes pictures, digitises every item that’s moving past on the conveyor belt.

‘It is then able to infer in real time what is a contaminant and what is recyclable. It then guides a robotic arm to pick and place the item in the correct bin.

‘Typically its about 80 picks a minute but it can get up to 120 picks and that is compared with a human who, best case scenario, is about 40 to 50, maybe 60, picks per minute.

‘And its allowed the people that work at these facilities to be shifted to different types of roles, that are focused on technical skills, maintenance and other critical roles. We’re letting the robots handle the real dirty parts of the process that are dangerous and a hazard for humans.’

Carling said that the technology has helped save ‘1.7 million metric tonnes of greenhouses gas emissions, roughly equivalent to removing about 360,000 cars from the road.’

Ms Spelhaug believes that the technology can help address the challenges the recycling industry faces.

Listen to the full podcast above and don’t forget to subscribe to our newsletter.

Make sure you check out the latest Net Hero Podcast episode:

Latest Podcast