Skip to content
Book Open access

Programmable Machine-Knitted Sensors for Deformable Interfaces

Nov 2026 · ACM Symposium on User Interface Software and Technology · 0 citations · 73 references
Computer Science

Abstract

Knitted textiles are well suited to wearable sensing because they are soft, conform to the body, and remain comfortable during wear. Industrial knitting also allows sensing layouts to be defined digitally. We present a programmable machine-knitted capacitive sensing substrate for wearable deformation sensing. The substrate realizes five functional layers through three physical knitted layers, with conductive yarn stripes and dielectric regions formed directly during knitting. The same sensing structure can be adapted to different wearable forms by modifying the knit pattern, which we demonstrate with elbow and shoulder prototypes. To evaluate how well the fabric captures deformation, we reconstruct the deforming fabric surface from spatial capacitance patterns using synchronized motion capture as supervision. We compare five model families under a six-fold leave-one-session-out protocol, with a Fourier Neural Operator (FNO) representing the field-based family. FNO and ConvLSTM achieve the lowest elbow reconstruction errors; on the shoulder, FNO has the lowest fabric-only mean and median error, while ConvLSTM does not improve over the single-frame baselines. These comparisons show that the relative value of spatial, temporal, and auxiliary pose representations depends on the wearable form. Surface reconstruction provides a shared geometric basis for evaluating programmable knitted deformation sensors across body regions.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.