I report on my empirical work which shows that effective human teams produced about twice as many self-repair disfluencies as ineffective teams in a collaborative search task. I then describe ongoing work to implement disfluency-handling mechanisms for robots that serve on teams with humans.
@inproceedings{gervits2017pioneers,
title={Disfluency Handling for Robot Teammates},
author={Felix Gervits},
year={2017},
booktitle={Proceedings of the Human-Robot Interaction (HRI) Pioneers Workshop},
url={https://hrilab.tufts.edu/publications/gervits2017pioneers.pdf}
}