speaker-info

Stefan Stoican

PhD Student, Essex

Biography

Stefan’s academic background may lie in Mathematics and Psychology, but his interdisciplinary mindset has constantly pushed him towards games and Computer Science. For his final Mathematics project, he designed an Android app that gamified teaching statistics. As part of his Psychology Masters degree, he investigated the potential benefits of MOBA games such as League of Legends with regard to visual attention. Currently, his extracurricular projects aim to explore video games’ effects on coping with trauma and on one’s perception of vulnerable groups, via commemorative gaming name choices or via in-game refugee storylines, respectively.

Thesis

Understanding human crowd behaviour via virtual environments: feedback loop between games & research

This project uses computer game experiments to explore decision-making in a virtual evacuation simulation.

Can one be “saved by the gaze”? Currently, Stefan is investigating how innate social cognition components such as gaze-cuing might inform one’s egress. Do “Us versus Them” scenarios occur? He is also testing how one’s feelings of social identification with the surrounding crowd might modulate one’s risk-taking.

Does hoarding prevent herding? Lastly, the project is looking at how cultural differences might affect egress time, when one insists to save personal possessions. More broadly, Stefan’s research concentrates on two key open questions in human crowd behavioural research. Firstly, how do social groups (that the player observes or is a member of) within the simulated crowd of agents affect both individual decision-making and the emergent behaviour of the crowd?  Secondly, both empirical and virtual experiments of human crowds have not fully explored the effect of agent or player interactions with underlying landscape features (e.g. layout, signage, debris, large objects and other obstacles, etc).

The outcomes of the experimental studies using real human participants will subsequently be used to develop more realistic decision-making and behavioural response algorithms and hence improve the behaviour of simulated agents in follow-on computer games.

Details

Home Institution: University of Essex
Supervisors: Edward Codling (Essex) and Dan Franks (York)
Email: sastoi@essex.ac.uk

My Sessions