Efficient Terrain Reasoning Algorithms for the Combined Arms Tactical Team Trainer (CATT)

Current-generation SAF terrain reasoning algorithms consume a great deal of computational resources. This doesn't leave enough resource, to run better, more sophisticated terrain reasoning algorithms or additional, more sophisticated higher-level behavioral models. This adversely effects fidelity and requires more expensive hardware. Emerging, more sophisticated terrain algorithms will exacerbate this problem. First, we propose to analyze, profile, optimize, and reevaluate some controversial design decisions of the latest terrain database libraries. Second, we propose to examine whether new, powerful commodity multiprocessors are able to provide more power to run current and future terrain reasoning algorithms. If so, we propose to develop new parallel algorithms, based on the current state of the art, that will make efficient use of these new multiprocessors. The resulting systems will achieve higher perfomance at lower costs and enable use of more sophisticated, higher-fidelity terrain reasoning and behavioral algorithms. We propose to capitalize on expertise obtained from: writing terrain database and reasoning libraries, developing entire SAF systems and components, and from carrying out two previous successful efforts to boost SAF performance. This experience leads us to propose analyzing the current SAF terrain libraries, profiling them to determine their perfomance, optimizing them where possible, and reevaluating some of their design assumptions in search of even greater performance gains. The algorithm will give CCTT the capability to scale up to CATT and STOW-level exercises, to achieve higher fidelity, to support both training and combat development and to help drive down recurring costs. The approach has uses including multiplayer gaming and collaborative work environments. It will spawn dual-use development tools for, and applications which incorporate, terrain reasoning. These products will "spin off" to commercial markets and "spin on" back to DoD markets.