CHARLES RIVER ANALYTICS, INC. — Department of Defense SBIR Phase II: ABSTRACT: Participants in Red Flag (RF) exercises must be familiar with specific procedur

Description

ABSTRACT: Participants in Red Flag (RF) exercises must be familiar with specific procedures, conditions, and training rules for range safety. Preparing non-US participants requires foreign travel of USAF personnel, which is expensive in money, time, and personnel resources. Considerable savings could be realized with a foreign deployable pre-exercise training environment, requiring no in-country travel/participation by USAF personnel. To meet this need, under a Phase I effort we designed and demonstrated the feasibility of a Simulation for Ubiquitous RF Instruction and Readying (SUREFIRE). In Phase II, we will: 1) define scenario creation and training environment requirements based on the curriculum developed by our subject matter experts and an analysis of RF training needs; 2) develop a graphical scenario creation tool that allows RF Subject Matter Experts (SME) to create and deploy training; 3) integrate an existing simulation with intelligent hierarchically composable agents capable of naturalistic verbal interaction with pilots; and 4) evaluate the effectiveness of Red Flag preparation using SUREFIRE. SUREFIRE will provide a deployable and tailorable tool capable of supporting pre-exercise trainingfor both foreign and US personnelin preparation for Red Flag exercises. BENEFIT: Voice-enabled interactive agents can play a role in a wide variety of domains, from automated telephone receptionists (where they are currently used frequently) to commanding unmanned robotic vehicles. By integrating cutting-edge speech and natural language processing capabilities into Charles River Analytics"user-friendly agent development environment, AgentWorks, we can broaden its market appeal. In addition, generalized tools for 1) rapid scenario generation can be created and are highly useful for customized training; and 2) real-time performance assessment feedback can support self-paced training in a variety of domains.