Architectural challenges on the analysis of human behaviour in synthetic environments

In this paper we review how human interaction models are designed for its use in synthetic environments , being these any kind of computer-generated environment , as in virtual reality, augmented reality or others. We present how human interaction techniques are modelled in synthetic environments relating to the self-representation of the user's body, the avatar, and the virtual objects present in the world, and we also propose alternatives on the design methods of that interaction models. We review the differences between in-lab and out-of-lab deployments of such synthetic environments and the problems that arise when we move those interaction models to actual applications with requirements such as mobility, massive use over the Internet, and implementations based on third party VR/AR platforms. By using the architecture of a reference framework for synthetic environments interaction models design, we present how the problems impact the overall design of the framework and propose some changes to adapt it to the new use cases presented. We also propose new lines of work to have a more industrial, reusable, cost-effective approach when designing new synthetic environments' interaction models from a domain specific point of view.