Rapidly Deployable Visualization for Training and Simulation in Urban Terrains

State-of-the-art visualization and simulation technologies can be characterized as cumbersome, difficult to deploy, immobile, lacking realism and visual fidelity, and costly in terms of infrastructure maintenance and initial hardware investment. As a result, systems based on specialized hardware typically involve low-resolution data rendered on a low-fidelity, display. This document outlines a research plan that will introduce scalable, commodity-based, immersive visualization systems to the Fort Knox, UAMBL that support the rendering of photo-realistic scenes. It seeks to bring together academic research with military and corporate applications to develop leading-edge solutions that advance the state of technology and build a high-tech community in the state of Kentucky. The visualization environments under development by researchers at the eniversity of Kentucky are addressing problems associated with current display technology with an alternative model referred to as a Metaverse display, that is easily/rapidly deployable, selfconfiguring, self-monitoring, scalable, and offers high fidelity display with photo-realistic content. As part of this project, we expect to extend our approach to address the unique problems of a photorealistic display environment for military visualization, training and simulation applications. The visualization research will include automatic calibration, resolution improving techniques, frequency multiplexing across overlapping projectors, and support of immersive video playback. We will investigate the problems related to integrating very-high resolution, real-world image data into the immersive experience. The use of real-world imagery supports user experience based on a virtual environment that more closely represents the real data on which it is based. Imagery and other data may be collected in a controlled environment, or garnered from sensors deployed in the field. In the second case there arc significant issues related to secure information delivery, multimedia protocols, and quality of services guarantees. We will study networking protocols that address these problems such as con cast in the context of information delivery to high-resolution immersive visualizations_ Ultimately, the project will deploy a next-generation Metaverse display at the Fort Knox UAMBL. The display will support Image-Based Rendering (IBR) for simulation and training within the environment. As a feasibility demonstration, a photo-realistic image-based database of objects from the domain of Military Operations in Urban Terrain (MOUT), will be acquired for rendering complex scenes on the display. The database will include dismounted soldiers, non-combatants, military and civilian vehicles, and ancillary objects commonly found in an urban setting. While visual simulation has been implemented effectively for training in large open environments involving "Force on Force" simulation there has been little attention and/or development in simulating environments that can provide the required fidelity for realism in the unique domain of Urban Terrains. The proposed research program is a starting point to a longer-term goal of the automatic and opportunistic acquisition of image-based models from mobile sensors, stationary surveillance networks, and aerial imagery to generate a photo-realistic immersive experience that is representative of remote, real-world activity and battlefield situation in near real-time.