TL;DR
- AI waste recognition is moving from a few equipped trucks to complete fleets. Only then do collectors and governments get data on their entire territory.
- Detection of hazardous materials and alerts in the cab are becoming standard, and measured waste data is finding its way into sustainability reporting.
- Rematics sees the refuse truck as a starting point. Other waste infrastructure and urban measurements are possible next steps.
At the end of 2024, the first refuse trucks were running with D4T. Today there are more than 70, and together they analyse around 1.75 million lifts per year. What will AI waste recognition look like in five years? These are the developments Rematics is working towards, with the conditions that come with them.
From a few trucks to complete fleets
Today a collector often equips part of its fleet. That produces a sample. With a complete fleet, you measure every street, every week. Then an inter-municipal authority can map sorting behaviour street by street and aim a campaign at the neighbourhoods that need it.
The pace depends on budgets and procurement procedures. Refuse trucks last a long time, and retrofit budgets at public organisations take time. An installation of less than 4 hours does lower the threshold.
Safety becomes standard
Lithium batteries are in more and more products and cause fires in refuse trucks, containers and sorting centres. Rematics wants to activate detection of hazardous materials on every equipped vehicle. Today the driver gets an audio signal. A visual alert on the on-board computer, through the integration with MOBA, is planned.
Fewer non-conformities, measured instead of estimated
In commercial waste, D4T today measures on average 40 to 50 per cent recyclable material. Rematics aims to bring that down to around 20 per cent within five years, together with customers. Technology alone will not achieve that. It takes collectors and municipalities that act on the data, with customer conversations and targeted campaigns.
How that reduction translates into CO2 can be read in 5,900 tonnes less CO2 per year. That figure still rests on an assumption today. The goal is to replace it with twelve months of measured curves per customer.
Waste data in sustainability reporting
Companies and public organisations increasingly have to report on their impact. Waste figures in those reports are often an estimate. Measured data per container makes a verifiable CO2 figure possible, which companies can use in their CSRD reporting. On the impact page you can see how Rematics already calculates today.
The visual inspection of commercial residual waste may also shift. Rematics wants an automatic report with image evidence to be accepted as formal proof of that inspection.
Beyond the refuse truck
"Waste is our starting point, but certainly not our end point," says Bram Peetermans, CEO of Rematics. He mentions underground containers, glass containers, organic waste and sorting belts as possible extensions of the same technology.
A refuse truck drives down every street every week. That makes it a logical platform for other measurements in the city, such as litter, the condition of the road surface, air quality or noise. These are possibilities, not existing products.
International
Since MOBA's investment, pilots have been running in Germany, Spain and the United Kingdom. Markets in South-East Asia and the Middle East are also being explored. Every new market needs adaptation: different fractions, different bags and containers, different rules. Read more about the partnership with MOBA.
What this requires
Three things determine how fast this goes: how quickly fleets are fully equipped, whether measured waste quality becomes a recognised indicator, and how regulation of AI in public spaces evolves. D4T processes waste contents and not people, but the European AI Act applies to the whole sector.
Want to see today what AI waste recognition means for your organisation? See how D4T works or request a demo.