The start-up Akanthas – “thorns” in Greek – has developed a digital technology whose objective is to help players in the waste sector to better manage this currently under-exploited resource. Based on image analysis using AI systems, the solution makes it possible to measure the filling rate of skips, but also to recognize the types of waste present in order to make the best use of it. Supported by IFP Énergies nouvelles (IFPEN), the start-up is now seeking to accelerate the development of its digital solution for managing and monitoring industrial waste.
A “Linky counter” for waste. This is, in essence, what the Toulouse start-up Akanthas, created in 2021, has developed. “Our goal was to digitize waste production. The solution we have developed makes it possible to monitor everything that is deposited in a dumpster. As we could do with a connected meter for electricity, water or gas, we come to capture the information, but also to transform and analyze it in order to allow the actors of the sector to make the best possible decisions »summarizes Viviana Contreras Moreno, CEO and co-founder of Akanthas.
A connected sensor
The solution is thus based on an intelligent – and autonomous – sensor allowing the operators of waste production, collection or treatment facilities to obtain information such as the filling rate of their skips, the nature of the waste deposited, or the possible presence of undesirable elements. A sensor developed by the start-up itself, as Viviana Contreras Moreno points out: “When initially testing equipment available off-the-shelf, we quickly realized that it was limiting, particularly in terms of energy efficiency. So we designed our own sensor.”
Like a camera, the device developed by Akanthas produces a timelapse of the contents of the dumpsters above which it is installed. Captured at a regular rate, the images are then analyzed using artificial intelligence algorithms. “We use several AI models. These models use images, but are also based on temporal data. They therefore allow an interpretation, but also a prediction.says Viviana Contreras Moreno.
Already high performance, and constantly improving
The device achieves an accuracy of around 90% when it comes to measuring the filling rate of skips. Material recognition is still in the calibration phase, as the co-founder of the company explains: “The system’s performance in this area is enhanced as the solution is deployed with different customers. The sensor already detects the main families of waste, but there is still room for improvement”. The Toulouse start-up turned to IFP Énergies nouvelles (IFPEN), a major player in research in the fields of energy, transport and the environment. The former French Petroleum Institute has in fact developed very specific expertise in the field of AI and image processing; expertise that IFPEN will therefore make available to Akanthas as part of its innovation support system for SMEs and start-ups. “This support is a real advantage for us, since it allows us to accelerate our developments around AI and the expansion of the types of materials detected”explains Viviana Contreras Moreno.
After wood, cardboard or even plastic, the start-up is working on the detection of more specific waste such as insulation or plasterboard. Enough to broaden the range of sectors for which its solution is intended. “We have a sectoral approach. We work both with professional waste producers, such as the manufacturing, agri-food, construction sector, etc., and with waste management operators, i.e. collectors and recyclers who take charge of this waste to recover it”, details Viviana Contreras Moreno. One year after its creation and the deployment of its first pilots, the company already has six customers. “We hope, by the end of the year, to exceed ten customers, which should involve monitoring two to three hundred containers”reveals the CEO of Akanthas.
A subscription offer
In terms of the economic model, Akanthas relies on subscription, avoiding hardware investments for its customers. The start-up thus offers a monthly package including the sensor as well as the data analysis services associated with it. A subscription which also includes the maintenance operations necessary for the proper functioning of the solution, and the amount of which varies according to the number of bins or waste storage cells equipped. “We offer a turnkey solution, very customizable according to the specific needs of the client”emphasizes Viviana Contreras Moreno.
And in addition to this monthly subscription, Akanthas offers a more punctual support offer, as its co-founder details: “We are also working in diagnostic mode. This is generally spread over eight weeks, during which we come to install the equipment and then monitor the activity in order to identify what is wrong and thus issue, thanks to our system and its AI, recommendations for concrete actions. . This makes it possible to advise the actors on the best practices, the most relevant actions to be implemented in the field”.
Convinced that waste is a resource far from being 100% exploited, the leader of the start-up intends, with her solution, to act directly at the source to make the most of it. “One in two dumpsters leaving industrial sites and construction sites sees part of its contents buried or incinerated while there is recoverable waste inside”affirms Viviana Contreras Moreno, underlining the financial, but also environmental consequences that this generates. “We come to act both on the financial, operational, but also environmental part, by allowing the development of more sustainable approaches, integrated in the circular economy”concludes the co-founder of Akanthas.
Enough to “remove a thorn from the side” of waste actors… “Waste is often seen as such by all stakeholders. Our objective is therefore to come and remove this “thorn”, hence the name of the company, Akanthas, which is the Greek translation”, finally slips the leader of the young company whose logo is none other than an owl with a piercing gaze. Another symbol used this time as a metaphor for the performance of its sensor in terms of vision.