Evaluating the Engagement with Social Robots

Abstract

To interact and cooperate with humans in their daily-life activities, robots should exhibit humanlike "intelligence". This skill will substantially emerge from the interconnection of all the algorithms used to ensure cognitive and interaction capabilities. While new robotics technologies allow us to extend such abilities, their evaluation for social interaction is still challenging. The quality of a human-robot interaction can not be reduced to the evaluation of the employed algorithms: we should integrate the engagement information that naturally arises during interaction in response to the robot's behaviors. In this paper we want to show a practical approach to evaluate the engagement aroused during interactions between humans and social robots. We will introduce a set of metrics useful in direct, face to face scenarios, based on the behaviors analysis of the human partners. We will show how such metrics are use

Multimodal Detection of Engagement in Groups of Children Using Rank Learning

“…The automatic detection of engagement using multi-modal cues has been investigated in the field of Human Robot Interaction (HRI) and Social Signal Processing (SSP) [35,38,4]. In [29], hand-coded features such as speech, gaze, gesture, and postures of two children were utilised to model individual and grouplevel engagement.…”

“…While the HRI studies [4,29] revealed that gaze-oriented movements and hand gestures were related to engagement, the settings often had regulations on behaviours of subjects or relied on hand-coded features. Without any regulation, the automatic extraction of these features is limited in naturalistic settings.…”

“…Fig. 2 presents each session's proportion of engagement types (the average proportions are 7.1 ± 4.7%, 40.9 ± 9.1%, (4), mean-duration pause (4) SD-duration speaker change (4), total-duration speaker change with overlap (4), total-count successful interruption (4), unsuccessful interruption (4), overlap (4) body (7) movement (4) mean-amount SD-amount total-count total-amount orientation (1) mean-orientation position x (1) mean-position position y (1) 7.9 ± 2.1%, 33.3 ± 11.8, and 10.8 ± 11.7%, respectively). Major portions are "oneless-engaged" and "all-different", which means that the children frequently exhibited different levels of engagement.…”

“…gaze, blinking, gesture) which have semantic information for engagement [1]. While these features were widely used in controlled or laboratory settings [29,4,6], we should investigate rigorous methods to extract these features in naturalistic settings where we cannot regulate children's behaviours.…”

“…The automatic analysis and detection of engagement have been studied before using multi-modal cues such as speech activity, gaze, posture, and gestures [29,6,19,21,4]. These features can be categorised into vocal and visual features, and each category has its own strengths and drawbacks in the wild.…”