Knitted Sensors: Designs and Novel Approaches for Real-Time, Real-World Sensing
summary
A digitally knitted capacitive fabric uses one conductive carbon-coated yarn with only two external connections to create a 36-button sensor and several form factors like a piano keyboard and volume pad. Real-time acquisition uses swept-frequency Bode analysis over 96 frequencies producing 250 temporally related measurements per touch, reduced to statistics and a baseline subtraction. A three-layer LSTM model trained with 386 samples per class yields 66% subject-independent accuracy (chance ≈3%), with EMI robustness at 80% and stretch robustness at 37%.
AI-written summary · about
chapters (9)
- 0:00Intro
- 0:41Effects of Design Simplicity
- 1:05Toward Real-Time, Real-World Sensing
- 1:33System Overview
- 2:09Real-Time Signal Generation and Processing
- 2:51Feature Construction
- 3:18Classification using Neural Networks
- 3:55User Studies . Model Training: data collection to learn the network
- 5:14Knitted Designs and Applications
related talks
18mRESi: A Highly Flexible, Pressure-Sensitive, Imperceptible Textile Interface Based on Resistive ...
18mStretchEBand: Enabling fabric-based interactions through Rapid Fabrication of Textile Stretch ...
28mWeaving the web - Programming textile-based interactions
16mPaintable IOT: Scalable sensing
19mElectrick: Low-Cost Touch Sensing on Arbitrary Surfaces Using Electric Field Sensing
18m“I don’t want to wear a screen”: Probing Perceptions Of and Possibilities for Dynamic Displays ...
33mBuilding "learn to touch type" glove with ...
31mHybrid Body Craft
