Marc Brunero

Interview with Marc Brunero

Max-AI®, the operator-sorter of the future

Marc Brunero, Direction Technique & Performance, Activité Recyclage & Valorisation des Déchets (RVD) de Veolia en France

Marc Brunero

 

For over twenty years, Veolia has been devising innovative technologies for sorting its clients’ waste more effectively. Its key innovations include remotely operated sorting — refining the sorting using touchscreens — and auto-adaptive sequential sorting (TSA2®) — automated sorting of packaging according to its material and color. The Group is now integrating smart sorter robots into its industrial facilities. Veolia’s Technique and Performance division, which has been supporting the Group’s sorting centers in their technological evolution, is banking on artificial intelligence. The Max-AI® robot — a first in France and Europe — has been operational in Amiens since June 2018.

What does artificial intelligence bring to high-performance sorting?

The artificial intelligence robotics solution supplements the other cutting-edge solutions that we use in our sorting centers. It fits in well with our primary goal of continuing to simplify the sorters’ actions, which enhances occupational safety while boosting the material flow sorting performance. We must deal with ever more ambitious sorting demands. On the one hand, expanding the waste-sorting guidelines, which is key in increasing the amount of recycled materials, has meant that separate collections contain waste that is more soiled than before and therefore more difficult to sort. On the other hand, China’s ban on low-quality plastic waste is forcing us to improve sorting quality, while increasing our performance to ensure an outlet for the raw materials from recycling. Hence the interest in developing robots with artificial intelligence that are capable of making over 3,000 movements an hour to carefully sort waste.

In concrete terms, what does Max-AI® look like?

Max-AI® is the combination of an “eye,” a simple optical camera, and an “arm,” an articulated robot, controlled by a “brain,” a neural network implanted in a computer.

How are the two Max-AI® prototypes used by Veolia performing?

The installation of Max-AI® is a first in France and Europe. As the first models came from the United States, you have to teach it everything. In other words, we have to sufficiently enrich its database so that it can carry out its tasks correctly in its new environment. We are in constant contact with the engineers who designed it and are developing the necessary applications. It is up to us to adapt this American robot to French norms and specifications. The first is in test configuration in the mechatronics area of the Hall in Mantes-La-Ville to fine-tune its performance and check its future adaptation capacities in our different sorting centers. We decided to set up the second one in production at the Amiens sorting center to check its robustness over time.

Is Max-AI® versatile?

Reinforcing its versatility is an issue for the future. Today, it is only operational on one application: cardboard quality control. However, in the long term, it will doubtless be able to perform all of an operator’s tasks. Its current limits are linked to its learning curve: it must gradually learn to separate each type of waste better. It is precisely to optimize the robot’s function and accurately test its efficacy on different material flows that we are working on a second prototype with the Group’s Research & Innovation department.

Is the Max-AI® family going to be expanded?

These first experiments allow us to promote Max-AI®’s deployment in calls for tender concerning sorting centers for local authorities. This is the case in Nantes, where the Waste Solutions proposal was chosen by Nantes Metropolis & La Carene to design and build a new sorting center. We are going to set up two new Max-AI®s there. They will be operational as of 2020.

More:

Automation at the heart of sorting recyclable packaging

Read the full content of the record month

Keywords

Waste recovery

For further