Effects of etiquette strategy on human-robot interaction in a simulated medicine delivery task

The objective of this study was to examine the extent to which a model of linguistic etiquette in human– human interaction could be applied to human–robot interaction (HRI) domain, and how different etiquette strategies proposed through the model might influence performance of humans and robots as mediated by manipulations of robot physical features, in a simulated medicine delivery task. A “wizard of Oz” experiment was conducted in which either a humanoid robot or a mechanical-looking robot was used to present medicine reminding utterances (following different etiquette strategies) to participants, who were engaged in a primary cognitive task (a Sudoku puzzle). Results revealed the etiquette model to partially extend to the HRI domain. Participants were not sensitive to positive language from robots (e.g., appreciation of human values/wants) and such a strategy did not succeed in supporting or enhancing the “positive face” of human users. Both “bald” (no linguistic courtesy) and mixed strategies (positive and “negative face” (minimizing user imposition) saving) resulted in moderate user perceived etiquette scores (PE).However, individual differences suggested such robot linguistic strategies should be applied with caution. Opposite to this, a negative face saving strategy (supporting user freedom of choice) promoted user task and robot performance (in terms of user response time to robot requests), and resulted in the highest PE score. There was also evidence that humanoid robot features provide additional social cues that may be used by patients and support human and robot performance, but not PE. These results provide a basis for determining appropriate etiquette strategies B. Zhu (B) · D. Kaber Edward P. Fitts Department of Industrial and Systems Engineering, North Carolina State University, Raleigh, NC, 27695-7906, USA e-mail: [email protected]; [email protected] D. Kaber e-mail: [email protected] and robot appearance to promote better collaborative task performances for future health care delivery applications of service robots.