SBIR-STTR Award

Image Generation systems today currently can not easily represent Digital Elevation Models (DEMs) with unique but important topographical features such as tunnels, overhangs, multi-level highways, etc. These types of features must be hand-crafted in an off-line process with the final static implementation loaded at run-time. All abstract information defining the feature as well as scalability is lost with this approach. In this SBIR, we propose a new approach where the abstraction of this feature data is preserved by creating the scene directly from the abstraction data itself. We will load this topographical feature source data during run-time along with the other source data directly to the scene. This underlying scene will be updated to allow dynamic interaction with these features during a training simulation exercise. We will use an XML-based approach to assemble various types of feature data that is loaded upon at run-time. This XML-based approach will be extended to support the existing military database re-use initiatives, such as NPSI, MSB, CDB and SE Core. This approach allows topographical feature source data to be changed and then re-visualized with no intermediate processing steps.

Keywords:
Abstraction, Spatial Access, Database Schema, Architecture, Visualization, Gpu-Based, Assembler, Topographical,

The objectives of the proposed effort for this SBIR is to develop abstract methods of defining mixed topographical features on rural and urban terrain and to develop optimized real-time rendering algorithms that allow dynamic interaction with the above complex features in a training simulation environment. Traditional approaches used to render complex objects depend heavily on the offline processing of abstract data in order to put it into a form that is acceptable to the chosen run-time system. There are fundamental disadvantages to this approach including loss of abstraction, increased I/O bandwidth requirements, reduced performance due to CPU-based algorithms, and an inability to efficiently and easily incorporate new data or modifications to existing data. While this approach was ideal in the context in which it was originally developed over 20 years ago, today it is being stretched to the limit of what is practical. This has resulted in the evolution of what was once a great improvement over previous approaches into what is now a typically very costly process for visualizing synthetic environments. The technology we will develop in Phase II will seek to address each of these shortcomings by employing an innovative approach to run-time dynamic database construction and rendering.

Keywords:
Source Data Visualization, Source Data Visualization, Gpu-Based Algorithms, Terrain Rendering, Visual Simulation, Dynamic Construction, Real-Time Graphics, Hyper-Elevation, Da