Management of the temporal conformity of a dynamic virtual environment
Éric Boivin
Denis Poussart (Supervisor)
Denis Laurendeau (Co-supervisor)
Problem: To be able to update a virtual environment (VE) scenario while it is being used is a most attractive concept. This would allow interruptions in the virtual experience to be avoided following the application of a modification in the VE. Several software strategies, such as the dynamic upload of components, have enabled the modification of a scenario to become an easily executable operation. However, these modifications must be monitored to a high degree to ensure the maintenance of conformity with the defined virtuality.
Motivation: The progress in the modelling of dynamic environments has given rise to several VE challenges. In addition to the reuse of models, any form of dynamism can have positive effects on the dynamic scenario, the consolidation of persistent VEs and the improvement of the accessibility and emergence of a profitable business model.
Approach: The development of VEs has been heavily laden with a past motivated by distribution. Initial motivations concentrated their actions on technology considerations which have changed significantly over time. The proposed project will be oriented towards a more conceptual approach where influences associated with technology and distribution will be more discrete.
Challenges: The classification of various forms of dynamism, the clarification of the dependence phenomenon between models of a VE, the maintenance of conformity with the defined virtuality and the gradual automation of the planning of the execution sequence are all necessary objectives when striving towards a dynamic VE. In this regard, the understanding of various sub-concepts associated with dynamism, such as the scale, the extensibility, the interoperability and the ability to build virtual models are unavoidable prerequisites.
Expected results: In order to restrict the research area, aspects related to distribution such as parallelization, load balancing, the addition of resources and remote users will not be dealt with. Moreover, finding a real-time solution is not a priority.
Calendar: 2002 to 2005
Last modification: 2003/04/24 by eboivin


©2002-. Computer Vision and Systems Laboratory. All rights reserved