ITR/SI: Real-Time Long-Distance Terascale Computation for Full Bandwidth Tele-Immersion

点击翻译按钮获取中文摘要

Tele-immersion, an ability to share presence with distant individuals, situations, and environments, may provide vast new possibilities for human experience and interaction in the near future. In order to make these dreams a reality, a host of technical difficulties must be solved: acquisition, communication, display, and interaction. This project focuses on answering two of the major questions: 1) Is tele-immersion as good as its promise or is there some inherent weakness? By using overwhelming computation power, the PI hopes to accelerate development so a real system can be tried within three years. 2) Is there a show-stopper with tele-immersion over long distances? The PI will try it for real, not just simulated distance via artificial delays. Specifically, the PI will build a 60-camera system at the University of Pennsylvania, compute depth on the 3000-processor Terascale Computing System at the Pittsburgh Supercomputing Center, and display at the University of North Carolina. The PI will have a more modest acquisition system at UNC (15 cameras) and a more modest display system at UPenn to enable local development and two-way collaboration. The testbed application will be remote medical consultation by a trauma surgeon, chosen because of its inherent demand for real-time response and its great potential benefit to societal. The PI expects to be able to provide demonstrable progress within the three year period of the project. Team members who are leading trauma surgeons will guide the design; evaluation will be by other trauma surgeons. The three years of effort will lead to a testable prototype with sufficient power and scope to be evaluated by application experts not on the research team. The results will indicate the likely success of other tele-immersion applications, as well as providing experience in building real-time applications on a remote terascale facility over fast wide-area networks.