Fig. 1: Knitted samples with temperature sensor fibers, Fig. 2: Pleated fabric with embedded Accelerometer sensor fiber and sensing training using different sand weights, Fig. 3 & 4: Edge Impulse NN training data & confusion matrix
Fig. 1, 2, 3 & 4: ExuTex fabric constructed as a sleeve, Fig. 5: Fabric composition, Fig. 6: Edge Impulse NN training data & confusion matrix
Computing Fabrics
This studio brought together machine learning, physical computing with Arduino, and advanced textile fabrication to investigate new possibilities for sensor-embedded fibers engineered for extreme environments. By integrating custom electronic yarns with computational design workflows, the project explored how intelligent textiles can sense, interpret, and respond to harsh contextual conditions.
Building on this foundation, the project focuses on designing a smart fabric, ExuTex that addresses the persistent challenge of lunar regolith—an abrasive, electrostatically charged dust that clings stubbornly to spacesuit surfaces. The first trial integrates accelerometer-based yarns and a single vibration motor; the second-final trial integrates a capacitive yarn within a lattice structure of nine vibration motors to detect localized dust accumulation through micro-weight changes and mechanically shake the particles free. By embedding sensing and actuation directly within the textile architecture, the fabric functions as an active self-cleaning surface, reducing contamination, improving mobility, and enhancing the long-term performance of spacesuit assemblies in extreme off-Earth environments.
Acknowledgements: This work was done in collaboration between fibers@MIT (Department of Material Science & Engineering) and RISD Textiles studio Computing Fabrics.