A Distributed Approach for Automatic Speed Adjustment during Navigation in 3D Multiscale Virtual Environments
Resumo
Multiscale virtual environments (MSVEs) are virtual environments that encapsulate elements with different scales within a same shared space. They may contain elements with extremely different levels of scale in the same environment and can be found in geographical maps and engineering virtual environments (VEs). Due to its complexity and diverging levels of scale, this type of environment could be accessed using its hierarchical structure to navigate between the levels of scale. But, in geographical maps and engineering VEs users may need to freely navigate through the levels of scale seamlessly to provide improved spatial knowledge. Common approaches to navigate within these 3D virtual environments are based on the Automatic Speed Adjustment approach, where the scene is pre-processed in a data structure and then accessed in real-time to determine optimal speed. A common trend is to migrate this process from the CPU to GPU, but the works that use this approach are still limited to static and/or smaller virtual environments. As the scene grows in complexity, the computational power needed to render the 3D scene and determine optimal speed may be too costly. We propose RMNS (Remote Multiscale Network System) to solve this problem by using a service-oriented approach to compute speed and is asynchronously accessed to determine optimal speed. Our results show that our approach enables the use of dynamic scenes, as the rendering and speed adjustment are decoupled. Finally, the asynchronous nature of our approach showed that the response time is enough to support automatic speed adjustment, while maintaining the rendering in real-time.
