Current simulation-based training systems typically support training for conventional combat operations against organized military threat and require large numbers of support personnel to operate. These systems model very limited types of Non Player Characters (NPCs) used to simulate threat forces and non-combatants. The cost and time required to expand the capability of these systems by adding new NPC types, to meet additional training needs, is prohibitive. The expense of these systems makes them largely unavailable to our military personnel at the frequency they require, where they need it. Finally, the time required to develop training scenarios is unacceptable, running into months, when even a few hours may be too long for out military personnel to wait. In this effort we will design and prototype an efficient, low-cost capability to quickly model large numbers of NPC types with the abilities to quickly input cultural aspects and traits. The resulting system will allow users to quickly create and assess NPCs that can be directly inserted into training environments that support warfighters in all echelons and across all simulation domains. The system will significantly reduce the number of personnel required to support simulation-based training events. The effort will also reduce simulation development time and cost by streamlining the NPC modeling process.
Keywords: Non Player Characters, Contact, Culture, Simulation, Multiplayer Game, Artificial Intelligence, Behavior Emulation, Neural Nets