3DKnITS: Three-dimensional Digital Knitting of Intelligent Textile Sensor for Activity Recognition and Biomechanical...

Area of Science:

• Materials Science
• Textile Engineering
• Wearable Technology

Background:

• Developing intelligent textiles requires integrating sensing capabilities with comfortable and scalable fabrication methods.
• Existing methods often face challenges in achieving seamless integration, personalization, and long-term stability.

Purpose of the Study:

• To present a novel approach for creating scalable, piezo-resistive matrix-based intelligent textiles using digital knitting.
• To engineer both electrical and mechanical properties of sensing textiles through yarn customization and thermoforming.
• To demonstrate the application of these textiles in personalized wearables and interfaces.

Main Methods:

• Utilizing digital flat-bed and circular knitting machines with functional and common yarns.
• Incorporating a melting fiber for thermoforming 3D piezo-resistive fabric structures that conform to the body.
• Developing a wireless hardware system and deep learning models (Convolutional Neural Networks) for data analysis.

Main Results:

• Fabrication of a 45x45 cm intelligent mat with 256 pressure-sensing pixels and a 3D form-fitted shoe with 96 sensing pixels.
• Achieved linear piezo-resistive sensitivity of 39.4 for loads up to 500 N.
• Real-time classification of 7 basic activities/exercises (99.6% accuracy) and 7 yoga poses (98.7% accuracy) using personalized CNN models.

Conclusions:

• Digital knitting offers a versatile platform for fabricating 2D to 3D intelligent textiles and wearables.
• Thermoformed piezo-resistive textiles provide robust structures with intimate interfacing for enhanced accuracy and comfort.
• The developed technology shows significant potential for applications in rehabilitation, sports science, wearables, and gaming.