This research presents ExuTex, a composite textile that senses environmental particulate contact and sheds it through embedded vibrotactile actuation. By integrating capacitive sensing, on-device classification and mechanical response within a single wearable substrate, the system closes a sense-decide-act loop at the level of the material itself. Trained across mineral sand and styrofoam, materials of markedly different density and dielectric response, the classifier detects particulate contact as a category, generalizing beyond the signature of any one substance. We describe the composite's architecture, its classification pipeline and its fabrication, and position the work against prior research in e-textile sensing, actuating textiles and embedded machine learning. Spacesuit regolith mitigation is proposed as the primary application, alongside a secondary space of industrial protective wear. The prototype demonstrates the loop under wear and the paper closes on what material-scale inference opens for self-maintaining textiles.
Acknowledgements: This work was done in collaboration between fibers@MIT (Department of Material Science & Engineering) and RISD Textiles studio Computing Fabrics.