MOCAV: A Multi-modal Human Comfort Database for Autonomous Vehicles

Autonomous vehicle (AV) technologies have thrived with the continuous research efforts from the academia and industry. As AV mileage rapidly accumulates through the prevailing of advanced driver assistance systems and various experimental robotaxi services, passenger comfort remains a barrier to the broader acceptance of AVs. To date, there has not been an openly available dataset focusing on human comfort in AVs. This research presents MOCAV — a Multi-mOdality human Comfort database for Autonomous Vehicles. MOCAV contains the self-reported real-time comfort levels in simulated AV journeys of 20 participants along with the physiological signals and vehicular behaviors during the journeys. The dataset has empowered several peer-reviewed publications about comfort in AVs topics, including identifying factors of human comfort in AVs and quantifying human comfort in AVs. The outcomes of these studies generated suggestions for the comfort-aware designs of AVs and provided quantitative tools to investigate human comfort in AVs. This paper aims to introduce and share the dataset¹ with the community to empower further research on human comfort in AVs.