Cognitive Agents for Simulation of Battlefield Airspace (CASBA)

Distributed simulation has the potential to promote more useful experiments and more effective joint exercises, with platforms from all branches of the services operating within the same virtual battlespace. The advantages of a virtual battlefield airspace could be significantly extended through the use of automated forces, or computer-generated forces (CGFs). The Army has already employed this technique to reduce manpower needed to staff an exercise and thus reduce costs. However, broader gains are achievable if the entities standing in for human role-players were more capable of human-like decision making and spoken interaction than is currently the case. Synthetic entities must be capable of understanding spoken and datalinked ATC (and other air traffic agency) directives and of responding appropriately, both behaviorally and verbally. These entities must exhibit appropriate interactions with air traffic agencies regardless of whether the controllers are human or synthetic. To accomplish this goal, we propose Cognitive Agents for Simulation of Battlefield Airspace (CASBA), which will (1) realistically model the decision-making, judgment, situational awareness, workload management, and communications of airborne entities in a battlefield airpsace; (2) provide robust, speaker-independent speech recognition and understandable and tactically-realistic speech synthesis; and (3) function reliably and interoperably in distributed simulation experiments.