Collecting large-scale, open-vocabulary activity data with wearables in the wild is difficult due to costly, imprecise labeling. Existing apps rely on labor-intensive self-reports, treating participants mainly as crowd labelers. We present Pebbl, a system for opportunistic crowdsensing framed as a tiny-habit tracker: users create if-this-then-that recipes whose audio-linked triggers (time/location) prompt just-in-time watch notifications; two-tap confirmations yield timestamped sensor windows. Implemented on Apple Watch + iPhone, Pebbl was evaluated in three studies. In a workshop (n=6), wearable-HAR researchers rated Pebbl more scalable, valid, and less burdensome than baselines. In a lab study (n=21), Pebbl achieved reliable execution records (recall=97.3%), strong user preference, and lowest compensation requirements. In a pilot in-the-wild study (n=8), Pebbl again showed preference and usable performance, demonstrating promise for large-scale deployment. By embedding labeling into habit tracking, Pebbl enables a sustainable, user-centered ecosystem for wearable sensing and inspires future crowdsensing system design.
Publication:
Wearable Opportunistic Crowdsensing for Open-Vocabulary Activity Data Collection Through User-Scheduled Trigger–Action Routines
Anonymous author(s)
*This paper is submitted to the CHI Conference on Human Factors in Computing Systems (CHI ’26). It is currently under review and temporarily open for application. People who see this page are kindly requested not to spread it.
Project Credits:
PI Lab at Tsinghua University, Department of Computer Science and Technology, directed by Yuntao Wang.
This project is funded by the Tsinghua University Student Research Training program.