Towards an Automated Model to Evaluate Collaboration Through Non-Verbal Interaction in Collaborative Virtual Environments

Related Work

Nonverbal behavior has been broadly studied in the real world and for the creation of artificial behavior in robots or animation (Breazeal, Kidd, Thomaz, Hoffman, & Berlin, 2005). However, there are few studies of the nonverbal cues people display in CVE through their avatars. In some cases, the developed CVE are focused on the automatic generation and scripting of nonverbal behaviors for autonomous agents; and in others on real-time interaction of human users with the primary goal to offer a tool that allows sending basic emotional nonverbal messages.

Guye-Vuillème et al. (1998) established the importance of non-verbal communication in face-to-face interaction and its conversion to an equivalent in virtual worlds, studying the advantages and disadvantages of complex embodiments. Using their Virtual Life Network (VLNET) they presented a solution that takes into account the practical limitations of input devices and social science aspects. Back then, their work exposed virtual environments as cold, dehumanized places, and with static avatars lacking emotions; while they stand out nonverbal communication as the most efficient way to communicate emotional content. About the use of avatars, Guye-Vuillème et al., (Guye-Vuillème, Capin, Pandzic, Thalmann, & Thalmann, 1998) recognized its importance as the means for the interaction with a virtual world and sensing various attributes of it. Also, that the avatar becomes more important in CVE because in addition, the avatar has crucial functions as: perception, localization, identification, visualization of others' interest focus, visualization of others' actions, and a social representation of self through the customization of the avatar.