Jin Hee (Heather) Kim, Kunpeng Huang, Simone White, Melissa Conroy, Cindy Hsin-Liu Kao
ACM Designing Interactive Systems (DIS) 2021
Best Paper Honorable Mention Award
PDF I DOI
Introduction
KnitDermis started by imagining a compliant tactile system that renders various tactile feedback without the help of rigid hardware (e.g., tactors). KnitDermis is designed to deliver a wide range of mechanotactile feedback on the wearer’s body. KnitDermis presents machine-knitted substrates, which integrate shape-memory alloy micro-springs into various shapes of knitted channels. As these small-scale actuators run through the substrates in various shapes, they deform the skin mechanically and stimulate different receptors. KnitDermis takes advantage of shaping techniques of machine knitting, which allow it to generate slim, stretchable, and versatile forms. Thanks to this, KnitDermis conforms to protruded joints and convex body locations. The types of tactile feedback KnitDermis introduces include: compression, pinching, brushing, and twisting.
Fabrication of KnitDermis
Our approach for a compliant system for tactile feedback is based on SMA micro-springs that contract when activated. We integrate the SMA micro-springs into channels into the knit substrates. As Joule heating heats the micro-springs, they contract and deform the knit channels. When parts of the substrates are grounded on (attached to) the skin, the contraction applies normal or tangential force to the skin, stimulating low-threshold receptors (e.g., Ruffini endings). Another design option is stimulating a sensitive touch receptor, through the deformation of the device itself.
Mixed-methods user study
Due to the compliant form factor of the devices and the absence of an end-effector, we were curious whether a user could notice the feedback and distinguish each from the rest. We conducted a mixed-method study, in which statistical analyses were performed to identify the relationship between the interface and response variables, namely, noticeability and distinguishability. We analyzed the data through a linear mixed effects model: fixed effects of the model were interface and gender, while participant ID was regarded as random effects. On the qualitative front, we invited users to semi-structured interviews. We took a deep dive into how users perceived and accepted the unconventional form factor of this not-so-usual tactile device. (Please see Sections 7.2 and 7.3 for full findings)
