From Textile Deadstock to Intelligent Circular Materials: The FIBROBOT Journey Begins!
Every year, Europe generates more than 7 million tonnes of textile waste, while large quantities of high-quality fabrics never even reach the market. Among these are industrial textile deadstock rolls, materials discarded due to overproduction, design changes, or production variability rather than poor quality. Despite representing an estimated 10–20% of textile manufacturing losses, most remain underused because they are difficult to sort, classify, and process efficiently. Their diversity in fibres, colours, patterns, and structures makes traditional recycling slow, labour-intensive, and economically challenging. As a result, many valuable materials are downcycled into low-value applications, or simply discarded, despite their potential to become new products.
FIBROBOT aims to foster a change. As an external pilot of the ROB4GREEN project, it explores how Artificial Intelligence, robotics, and computational design can work together to unlock the hidden value of textile deadstocks. Instead of treating material variability as a problem, FIBROBOT turns it into an opportunity, using intelligent technologies to understand, sort, and transform textile waste into new engineered materials for interior and fashion applications.
Over the coming months, the project will develop and validate an innovative robotic workflow that combines AI-driven material recognition, adaptive robotic handling, traceability, and human-centred design. By demonstrating these technologies in a real industrial environment, FIBROBOT aims to show how digital innovation can help make textile manufacturing more circular, resource-efficient, and ready for the future.
What if pre-consumer textile waste became the starting point for innovation?
FIBROBOT introduces a new approach to textile manufacturing, where material intelligence, collaborative robotic, and design-driven decision-making work together to retain the value of industrial textile deadstock and enable scalable circular production.
From textile deadstock to engineered materials. Discover how FIBROBOT combines AI, robotics, and design to shape the future of circular manufacturing.
FIBROBOT is organised as a manufacturing architecture consisting of three interconnected processing stations (ST) supported by a Computational Design and Process Intelligence Layer. Together, these components combine AI, computer vision, robotics, and computational design to analyse, process, and repurpose industrial textile deadstock within a traceable, adaptive, and data-driven manufacturing workflow.
The process begins with ST1 – Textile Waste Sorting and Analysis, where AI algorithms combine production data with computer vision to identify fibre composition, textile structures, colour distribution, and pattern characteristics. Based on this analysis, the system supports intelligent sorting decisions, while robotic manipulators automatically handle and separate textile sections for further processing. Once mechanically processed into fibres, the materials enter ST2 – Fiber Characterisation and Physical & Digital Archiving. Here, RGB vision systems analyse colour consistency and material characteristics, storing it in a structured database aligned with Digital Product Passport (DPP) principles. This continuously updated knowledge base supports both material traceability and the optimisation of downstream manufacturing processes. In ST3 – Fiber Recomposition, adaptive robotic systems transform recovered fibres into engineered nonwoven materials through controlled fibre deposition, layering, and robotic needle felting. Guided by computational design tools, the system generates different material configurations for interior and fashion applications. AI-supported quality verification continuously compares the manufactured output with the intended design, enabling process optimisation and ensuring both aesthetic and functional consistency. Human operators remain actively involved through Human–Machine Interfaces (HMI), supervising operations and validating system decisions.
The Computational Design and Process Intelligence Layer connects all three stations by integrating material data, robotic operations, quality monitoring, and design information into a shared digital ecosystem. This continuous exchange of information supports data-driven decision-making, adaptive manufacturing, and full process traceability.
By integrating material intelligence, robotic automation, and human expertise, FIBROBOT aims to increase material recovery, reduce repetitive manual operations, improve worker safety, enhance process consistency, and demonstrate a scalable model for circular textile manufacturing that can be transferred to future industrial applications.
Where AI, robotics, and design turn textile deadstock into tomorrow’s circular materials.
Erminia D'Itria
Erminia D'Itria is Assistant Professor at the Department of Design, Politecnico di Milano, and a member of the Politecnico di Milano team in the FIBROBOT project. Her research focuses on circular design for fashion and textile systems, exploring design strategies, methods, and innovation approaches that support the transition towards a circular economy.