MUST: Smartwatch-based Multimodal Framework for Predicting Driver State and Takeover Performance

Status: Accepted, to appear in the Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems (CHI 2026).

Authors: Seokyong Sheem, Yoonji Cho, In Kwon Lee, Hyeonsu Cho, Taewoo Kim, Byeong Hak Kim, and Dohyeun Kim.

Contribution: Introduces smartwatch-based multimodal prediction of driver state and takeover performance for conditionally automated driving.

Links

• Project page
• Paper, DOI, ACM Digital Library, slides, talk, poster, and preprint links will be added here as they become available.

Overview

MUST is a smartwatch-based multimodal framework for predicting driver state and takeover performance in conditionally autonomous driving.

This work studies how unobtrusive wearable sensing and multimodal fusion can support reliable readiness assessment without intrusive hardware or fragile vision systems.

BibTeX

@inproceedings{sheem2026must,
  author = {Sheem, Seokyong and Cho, Yoonji and Lee, In Kwon and Cho, Hyeonsu and Kim, Taewoo and Kim, Byeong Hak and Kim, Dohyeun},
  title = {MUST: Smartwatch-based Multimodal Framework for Predicting Driver State and Takeover Performance},
  booktitle = {Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems},
  year = {2026}
}

Recommended citation: Sheem, S., Cho, Y., Lee, I. K., Cho, H., Kim, T., Kim, B. H., and Kim, D. (2026). "MUST: Smartwatch-based Multimodal Framework for Predicting Driver State and Takeover Performance." Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems.
Download Paper