Turn-taking, feedback and joint attention in situated human-robot interaction

Both lexical choice and prosody in users' feedback are correlated with uncertainty.

“In a study on human-robot interaction, Skantze et al. (2014) found that both the lexical choice and prosody in the users' feedback are correlated with uncertainty, and built a logistic regression classifier to combine these features.”

The authors show that gazing on the robot has a positive correlation with certainty.

“The authors show that gazing on the robot has a positive correlation with certainty, i.e., that users look more at the robot when they are certain of the path they have just taken through the map.”

Skantze et al. (2014) showed that the performance of participants in the map task was connected to the user's utterance timing and prosody, and that this was more important than the linguistic contents of their speech.

“They showed that the performance of participants in the map task was connected to the user's utterance timing and prosody, and that this was more important than the linguistic contents of their speech.”

Skantze et al. (2014) show that gaze can be used as a measure of uncertainty based on how it is used together with other feedback signals.

“Skantze et al. (2014) show that gaze can be used as a measure of uncertainty based on how it is used together with other feedback signals.”

Skantze et al. (2014) showed that a physical robot with a face could use its gaze to elicit feedback by either looking at the task or the user.

“This was elaborated upon by Skantze et al. (2014), who showed that a physical robot with a face could use its gaze to elicit feedback by either looking at the task (a map game) or the user.”