AI models for waste sourcing: RECONSTRUCT’s first steps
News releases
RECONSTRUCT has taken its first actions towards developing AI-based sourcing models for collecting and characterizing waste, but for these to be designed, testing and data on industrial waste sites are needed.
That is why, during an experimental campaign led by UNIVPM (Università Politecnica delle Marche, Italy), two visits were made at the beginning of October to two areas: a rehabilitation ground in Barcelona and a plant in Tarragona, Spain. These sites are overseen respectively by COMSA corporation, one of the project’s partners specialised in infrastructure development and industrial engineering, and Sorigué, another project partner whose expertise is in urban engineering, green spaces and water infrastructures.
Partners from ITeC, Brunel University of London, Iris Technology Solutions, Symbiosis Industrial and ICCS toured the facilities together, being the partners working on the collection, design and validatation of these AI-models.
During the visit, tests in situ were carried out. Now, the engineering process of these identification tools will soon begin, with the participation of all the institutions involved in these activities.
“Experimental tests were pivotal to properly design future long-term installations of sensors in areas of interest for the management and sorting operations of construction and demolition wastes, as well as to design Artificial Intelligence models for CDWs quantification and characterization.”
said Gloria Cosoli, post doctorate in Industrial Engineering and Mathematical Sciences for UNIVPM.
Sourcing waste correctly is not only necessary to ensure the best reuse of materials that would be otherwise discarded but also to identify and locate where these resources can be acquired, so better pathways for circularity in the industry can be found.