AI has many potential applications on construction sites. One project is tracking waste being thrown away on sites to inform ways of improving recycling and saving businesses money
In the Australian Bureau of Statistics’ most recent report on waste generation in Australia, construction was the industry found to be spending the most on waste services – a figure of $2 billion in the 2018–19 financial year. In that period 12.7 million tonnes of construction waste was estimated to have been produced – an increase of 22 per cent in just two years.
And this isn’t just an issue in Australia. Worldwide, around 30 per cent of solid waste produced is construction and demolition waste, with significant amounts of potentially recyclable material being buried in landfills.
At Northumbria University in the UK, a research team led by Dr Pablo Martinez Rodriguez is looking into how artificial intelligence (AI) could be used to better understand what waste is being generated and how construction companies could reduce their waste disposal costs.
Barriers to recycling
The Northumbria University team has recently secured £250,000 (A$396,478) in funding for its development of an AI-driven tool that scans waste as it is being thrown into skips on construction sites.
Speaking with Earthmovers & Excavators, Dr. Martinez Rodriguez says that increasing sustainability in the construction industry is a necessary part of addressing global environmental challenges, but in order to enact meaningful change, current waste generation needs to be better understood.
“There are several barriers to increasing the amount of recycling taking place in construction,” he says.
“The first one is that we know very little about what waste is generated on sites and how it is being handled, but at the moment there is little value for a company to separate out its waste. Therefore, most of construction waste ends up being mixed, which requires it to be segregated at a later stage.
“For example, for plastics, there’s very little value to it being separated from brick waste or ceramic waste. Certain adhesives present in brick waste can also render certain plastics unrecyclable. So, currently, most waste is thrown away into the same skip and then taken to landfill.
“The second barrier is that, in order for something to be recycled, there needs to be some kind of added value that compels companies to put additional processes in place to keep waste separate and recycle it. Currently there’s very little beyond sustainability. So, we’re trying to try to address that barrier by understanding the value of that product [in the context of circular economy] and how much waste is actually generated on sites.
“Implementing segregation for a particular waste product on a construction site is nothing new – they do that for plasterboard, glass and hazardous materials, so this wouldn’t be any different. There just needs to be a reason for companies to do it.”
Using AI
The project will develop advanced AI-driven decision-support systems to help construction managers monitor waste generation points, implement effective handling strategies, and assess project sustainability through new key performance indicators (KPIs). These KPIs will measure waste handling efficiency, resource utilisation, and adherence to sustainable practices.
“A camera is placed within the waste collection point on a site, and it will determine in real time how much there is of a specific waste product for all the different classes of waste that you would have on the construction site,” Martinez Rodriguez says.
“This then gives data on the waste composition coming out of the site, so 15 per cent ceramic waste, 12 per cent glass, 20 per cent brick, 20 per cent insulation and so on.
“What we’re trying to do with the decision support system is use that data to extrapolate that a particular day has seen more waste being disposed of, and of a particular type. That can then be associated with schedules or events that may have happened during the project that can help explain why that waste was generated.
“If you’re expecting some framing work to happen, and you’re expecting a lot of wood to be generated, but somehow that day just saw ceramic and brick waste generated, the business can look into why that happened. For example, is there a delay of some sort that has yet to be communicated? Unexpected work?
“So, the data generated by looking at the waste can be of value to other processes on the site, helping support decision making.
“In the end, what we aiming to do is try to understand how waste is being generated and how it relates to all the other activities that happen on site. We want to use it as a guide to reduce waste, obviously, but also to understand how waste actually interacts with other activities on those sites.”
Two cameras are currently capturing data on a construction site in the UK, which will help train the AI tool to better recognise different materials and quantify waste volumes. The next phase will be to install cameras in the waste collection point of a site for the entire period of a project.
“What we really want to have a much more value-oriented approach to waste, by saying to a business – millions of pounds worth of material is being thrown away by your company every year, is there something that you can do internally with that?” Martinez Rodriguez says.
“Eventually, we want to get to the point where there are zero waste construction sites.”
Wider applicability
The AI tool was originally developed to study municipal solid waste to aid in sorting waste on conveyor belts. Martinez Rodriguez says that the focus moved to the construction industry as such significant amounts of waste was being produced and sent to landfill, but the camera system could be applied to any waste-collection process in different industries, such as agriculture and manufacturing.
He says that increasing familiarity with AI tools such as ChatGPT is helping to open up a greater acceptance of AI, with construction already seeing it being applied in machine monitoring for example. However, despite there being a wide range of opportunities for AI applications within construction, there is still some discomfort around the installation of cameras on sites that the team has to address.
“There was reticence at the beginning around introducing cameras on construction sites, although we are not recording and we are complying with data protection laws,” Martinez Rodriguez says.
“But, we did some work with the site supervisor and staff to explain what was happening, and that we weren’t tracking people on the site, just the waste being thrown away, which eased concerns about the use of AI.
“There are a wide range of potential applications. You can do scheduling with AI, or anything that talks to the BIM model. You can use it for quality, health and safety, monitoring project progress, scanning. You can use it for anything that captures data.
“At the moment construction is still not the best industry for capturing data, but we’re getting there.
“There’s a lot more to gain in recycling material, from a company’s perspective. Even if we target no-cost waste management, so not focusing on making profit, that would benefit businesses. Plus, having these resources circle back into the economy rather than going to landfill brings significant advantages.”
For more information on this research, you can contact Dr Pablo Martinez Rodriguez at pablo.rodriguez@northumbria.ac.uk.
- Subscribe to our free weekly newsletter to receive the latest news in the earthmoving industry
- Never miss a great deal and subscribe to our monthly magazine
- Download a free copy of our latest digital magazine to catch up on the biggest news and developments in the earthmoving industry