Making recycling smarter: TOMRA and AI
In the ever-evolving world of technology, a notable shift is capturing the attention of industries worldwide. AI is reshaping our understanding of technology and redefining the way we approach our daily tasks and long-term projects.
TOMRA AI Summit 2023
On 28-29 November, TOMRA hosted in Oslo a company-wide AI Summit, bringing together 85 TOMRA team members for talks and workshops on AI. The event consisted of two days of talks and a collaborative workshop, with a goal to inform, inspire and interact. The agenda spanned infrastructure, recognition, machine learning and generative AI, and included guest speakers from TOMRA technology partners Microsoft, Google and PolyPerception.
AI technology is already playing an important role in many industries, and this will only increase in the future. This technology will positively contribute to our mission to Lead the Resource Revolution.
Applications for artificial intelligence in TOMRA - an AI driven reverse vending machine
Reverse vending machines automate the return of beverage containers (such as glass and plastic bottles, steel and aluminum cans, and cartons) for recycling in a deposit return scheme, as part of waste management strategies to combat plastic pollution, boost recycling rates, increase the supply of recyclable materials, change recycling habits, and build a circular economy.
At 2023’s EuroShop, the world’s largest retail trade fair, TOMRA previewed reverse vending concepts for recycling, utilizing AI. The “over-the-air" digital concept, with dashboards powered by AI, brought together different smart reverse vending machine data and functionalities – including notifications about action needed at the machine (such as bin emptying), information on cleaning and usage to better plan day-to-day operations, and self-configuration options.
A second, “smart service” concept at EuroShop used AI to ensure maximum machine uptime and avoid unplanned repairs during recycling. It shows retailers the health of reverse vending machine components, and proactively notifies operators about action needed or servicing activities under way. The concept focused on predictive service and maintenance, using machine learning to help TOMRA service technicians see, for example, when the compactor (for crushing beverage containers in a reverse vending machine) would need to be changed. This seeks to extend the lifetime of components, make recycling operations more efficient, and give peace of mind for operators.
TOMRA’s other divisions are also exploring the applications of AI.
- TOMRA Food: In food sorting, AI provides opportunities to detect more defects than are visible to the human eye or traditional optical technologies, leading to less waste, higher yield and increased profitability for food growers and processors. TOMRA Food first deployed deep-learning models in early 2019 with LUCAi sorters, and in 2023 released new solutions powered by deep learning and machine learning.
- TOMRA Recycling: AI is incorporated in TOMRA’s sensor-based sorting machines for mixed waste, metal and wood sorting. In the past few years, TOMRA Recycling has invested in deep-learning-based technology, which allows facility operators to detect and recover materials that cannot be sorted with traditional technology. The division already has some deep-learning applications launched and will further invest in developing additional solutions in the near future.
“AI is not new to TOMRA. We are one of the few companies that have already used and commercialized deep-learning algorithms as part of our sorters, both for food and recycling. There it’s used as a complement and enhancement to our existing machines,” explained Tove Andersen.
There are a lot of different opportunities for AI. Imagine a tool that can continuously learn, adapt, and produce results tailored to our specific needs. That's the potential generative AI holds, also for our organization.
TOMRA’s principles for the exploration of AI
As interest around AI in business and technology accelerates, discussions around the guiding principles become critically important.
“TOMRA wants to ensure we develop and input data into AI systems in a closed and secure environment, thoroughly examining how we can use AI in our recycling with good policies and compliance, while also protecting intellectual property,” explained Mickaël Fontaine.
“Everything that is visible to us as humans, a machine can be trained to detect with a camera. This creates great opportunities. We need to be open to thinking differently, about new ways to understand needs and develop solutions,” said Tove Andersen. “I believe TOMRA needs to take a positive role to ensure the technology is deployed in a way that contributes to our mission to Lead the Resource Revolution.”
The path ahead for AI in TOMRA Collection
TOMRA has for the past years been working on the topic of AI, machine learning and deep learning, and will continue to focus on it as an important area for future development and competence in the company, and possibilities for an AI driven reverse vending machine.
“AI is a great opportunity to enhance reverse vending machines for all stakeholders. We are excited to host the TOMRA AI Summit, bringing together so many eager and bright minds, and truly building collaboration across the company in this integral area,” said Mickaël Fontaine.
“TOMRA Collection is looking at AI’s applications for product development both for evolution of our current reverse vending machines portfolio and a true revolution of our future work – while developing with good policies for safety and compliance.”
We look forward to working further on AI and delivering on its potential to bring an even greater drink container recycling experience for consumers and operators.
Learn more about TOMRA's reverse vending technology
What’s inside reverse vending machines?
Sensor technology in reverse vending
