A pioneering system to facilitate clothes recycling

UPC and CARNET create a prototype of an autonomous vehicle for delivering goods that will be tested in Esplugues de Llobregat
January 25, 2022
ARIADNA draws to a close: a European project to promote the use of GALILEO satellite technology applied to urban mobility
January 17, 2022
July 2021

The Motion Control and Industrial Applications (MCIA UPC) centre has taken part in the AUTSORT (Automatic Textile Sorting) project. As part of this project, a pioneering system has been developed to sort textile materials depending on the fibres that they contain, based on near-infrared (NIR) spectroscopy technology.

In Europe, it is calculated that every person consumes an average of 26 kg of textile products every year. Consequently, one of the main challenges the textile sector faces is recycling, as most textiles are made using different threads that require very different recycling processes. In addition, European Directive 2018/851 obliges Member States to implement selective collection of textiles before 1 January 2025, so that textiles can be reintroduced into the chain of recycling to reduce final waste. The aim is to cut the amount of textile waste that ends up in dumps or incinerators, boost recycling in the textile industry, and reduce the carbon footprint of the production of new fibres, as recovered fibres can be reused.

Therefore, classifying clothes according to their composition is key to be able to recycle them and move from the current linear system to a circular system. Today, this classification is mainly carried out by hand and almost everything goes to the dump or the incinerator. Unlike the textile classification technologies that are on the market and are still in very early stages, AUTSORT includes sensors that can recognise NIR spectra with great precision (above 90–95% depending on the type of textile) and determine the presence of cotton, linen, silk, wool, polyester, polyamide, viscose and up to 13 other fractions of fibres, which enables their automatic classification.

For its implementation, this system should be developed so that it can be scaled up to industrial level and therefore integrated into sorting machinery that already exists and newly designed machines. To achieve this, powerful mathematical models must be developed (based on multivariate statistical analysis and convolutional neural networks) that enable NIR spectra to be processed and textile fibres to be classified into different groups.

The AUTSORT project was led by the MODACC cluster. Other collaborators were the Foundation for Textile Innovation (FITEX) and the University of Lleida. The project was financed by the Innovative Business Groups (AEI) programme of the Spanish Government’s Ministry of the Economy, Treasury and Competitiveness. The project lasted 9 months and had a budget of 87,000 euros.

Related Projects

This site is registered on wpml.org as a development site.