Alternative Reality, a new approach to VR

Speaker: Marc Cavazza

22nd September 2004 , 2pm , Devonshire G21/G22 Conference Room

Abstract

Most research work in Virtual Reality (VR) endeavours to match the highest possible standards of physical realism, whether in visual appearance, physical behaviour of objects, or user interaction. However, early ideas in VR were advocating a psychedelic departure from our everyday experience. The VR pioneers were in pursuit of original experiences and fantasy worlds, far from our common sense laws of Physics. Our research in Alternative Reality investigates the behavioural mechanisms by which it is possible to create immersive virtual worlds, in which laws of Physics could be redefined on a principled basis, so as to induce new experiences. From a technical perspective, this approach substitutes itself to traditional “Physics engines”, taking advantage of the fact that these operate on a discretisation of virtual world events. This discretisation makes it possible to integrate AI-based simulation mechanisms. We have targeted two elements of reality. The first consists of process-based mechanisms (such as fluid flows, heat exchange, etc.) which we have modelled using Qualitative Physics. The second one, causal perception, lies at the interface between perception and cognition, and plays an important role in our mechanistic explanation of the world. We will show how the artificial creation of co-occurrences between events can induce causal perception in subjects. We have developed a complete environment using a game engine for visualisation (because of its sophisticated event system) which has been adapted to immersive VR using a CAVE™-like hardware (the SAS Cube™) and the CaveUT™ software. In this talk, we will give various examples of such behaviours in different contexts, from fantasy worlds created as part of artistic briefs to everyday settings which can be used to explore subjects’ perception. This work, which was originally developed in the context of VR Arts, is also opening new research directions, such as the development of virtual environments supporting cognitive experiments, or new techniques to optimise objects’ behaviour in VR.