Research

WIT Lab designs interactive systems and interaction techniques for mobile, wearable, and embedded devices. Our work reflects the unique form factors and use contexts of these systems: they are miniaturized, handheld, wearable, ubiquitous, and operated when mobile, distracted, busy, or multi-tasking. Our designs are informed through empirical work that acquires a nuanced understanding of the capabilities and abilities of the human mind and body. They are developed using a combination of digital fabrication, hardware and software prototyping, and advanced analytic techniques such as computational optimization and deep learning. The goal of our work is to create interactive systems that are seamless, transparent, expressive, and effective. The primary current focus is on applying the modalities of touch, gaze, and motion to emerging device form factors such as smartwatches and smartglasses in critical wearable use scenarios such as augmented and virtual reality.

Check below for a brief introduction to some of our projects, or head over to the publications page to see full details

Touch Sensing Fingernails

Can we use our fingernails as a touch surface to support HMD use? We develop a set of five touch-sensitive finger nails and present a series of studies and evaluations of their potential to support interactive experiences: a design workshop to brainstorm interaction ideas; a qualitative study that captures the comfort of a wide range of touches - 144 in total; and two studies that examine the speed and stability with which users can make touches to their nails. We conclude with recommendations for a final set of 29 viable touches that are comfortable and can be performed accurately and rapidly.

Bioacoustic Authentication

How can we quickly and reliably authenticate to smartwatches? This work explores how audio signals transmitted through the wrist can serve as a unique signal for unlocking and securing a smartwatch. We developed custom surface conductance speaker and microphone systems to transmit and record through body audio and conducted extensive empirical work characterizing their performance. We show that good performance can be achieved by sampling signals while users make simple gestures such as a fist or open hand.

Smart Watch Tapping Gestures

How can we quickly and easily control smart watches? This work explores how rapid patterns of two-finger taps can be used to issue commands on a smart watch. The goal of the this work is to design interfaces that give access to a wide range of functionality without requiring users to navigate through menus or multiple screens of information.

Detecting Emotions Online

How do you feel online? Emotions are a defining aspect of the social media experience. We believe that digital interactions like clicking and typing can serve as digital biomarkers for affective states and have the potential to improve self-awareness of affective well-being. We explore these ideas in the context of social media use in studies that capture touch, motion, and eye-gaze data on a mobile phone while watching media clips and browsing Facebook feeds. We predict self-reported binary affect with an accuracy of up to 94%, suggesting that future social media services could be designed to flag and highlight the affective state of their users.