Easiness of acceptance metric for effective human robot interactions in therapeutic pet robots

Abstract

It is essential for robots to be highly usable to achieve increased acceptance among target population. In this paper, we put forward a metric that measures easiness of acceptance using the length of time to initiate interaction, total interaction time, time needed to identify perception-reaction phenomenon, its occurrence, number of violations and affective interaction factor. This measure would be useful as a parameter to compare different target groups, robots, and different nature of interaction. Experiments were conducted to validate the proposed metric in an eldercare nursing home using a therapeutic pet robot ERIC (Elderly Rehabilitation Interactive Companion), a dog-like robot. Results of the statistical analysis showed no significant difference in Easiness of Acceptance (EOA) estimation for between subject experiments in groups with similar characteristics for pre and post training trials. But, significant differences in EOA estimation was witnessed for within subject experiments for pre and post training trials.