SILVA, C. J.; http://lattes.cnpq.br/9515351024505096; SILVA, Carlota Joaquina e.
Resumen:
For ethical and safety reasons, studying the behaviour of people in emergency situations would be unfeasible using the same traditional field study methodology. Virtual Reality (VR) is a tool that makes it possible to assess these types of situations efficiently without putting the participant in danger and to collect data with greater accuracy, as the participant reacts realistically. Works in the literature have evaluated the adequacy of this tool for safety training in large public spaces and people's reaction in relation to emergency situations to verify compliance or not with safety signs. There are few works that address these issues in industrial installations and in the context of emergency situations with danger to the life and physical integrity of workers. This work will explore VR as a tool that will substantially contribute to the analysis of human behaviour in critical circumstances and to obtain more realistic data in the context under study. The main objective is to develop a methodology to study the effectiveness of safety systems for the evacuation of industrial installations in the face of social influence in emergency situations, using VR. The methodology was developed in the form of directed studies and prototypes using VR, divided into two stages, with Experiment A evaluating the feasibility of VR as a study tool and ascertaining the minimum number of virtual agents that could influence participants in emergency situations. The results of Experiment A indicated that the greater the number of virtual agents, the greater the social influence on the decision-making of the participant. Thus, given the results obtained in this study, a new methodology was proposed, called Experiment B, which evaluated the social influence in relation to the effect of crowds in emergency situations. Simultaneously, it is also proposed how VR interaction interfaces can improve the level of presence in the context in VR studies.
