PREFLAG: Preparing Allied Forces for Red Flag with Desktop Simulation and Speech-Interactive Agents
Small Business Information
1035 Virginia Drive, Suite 300, Fort Washington, PA, 19034
Director of Contracts
Director of Contracts
AbstractA typical Red Flag may involve 1,500 personnel and 100 aircraft representing four or five nations. Each event presents critical safety-of-flight and range considerations. Many participants are non-native speakers of English, in a context where technical, rapid, and information-rich communication can pose steep challenges even to native speakers; the potential for misunderstandings, miscommunications, and even mishaps, is thus even greater for participants for whom English is a second language. While the Air Force deploys squadrons to help participating nations prepare, this practice is expensive and demanding at a time when flying hours are a precious commodity. Needed is effective, pre-exercise orientation to the environment, range rules, safety-of-flight, communications protocols and local procedures and restrictions. Such training must provide realistic practice in spoken radio communication, in English, and must be easily distributed and readily updatable. To meet these complex challenges we will create an interactive desktop simulation called Preparing for Red Flag: Local, Air and Ground (PREFLAG). PREFLAG will allow users to arrive at Red Flag fully-prepared to follow the guidelines and instructions of the host units and qualified to communicate using FAA and USAF approved pilot-controller terminology, while reducing the need for USAF squadron deployments to assist participating nations. BENEFIT: Successful accomplishment of PREFLAG will represent a significant advance in how allied forces are prepared for Red Flag, resulting in safer, more effective exercises and improved coalition readiness. The benefits of this program though extend well beyond Red Flag. The technology developed under this program will provide student pilots, controllers, navigators and tactical crew with an indispensable tool for enriching their training with guided simulation focused on critical skills. The technology will enable automated tutors to provide anytime, on-demand training, performance assessment and remediation. Commercial applications can bring this same capability to aviation, maritime and land-based training for transportation, homeland security, law enforcement and other tactical team domains. A successful Phase I effort will facilitate future research and development in four areas: 1. Intelligent Tutoring - This work explores the use of intelligent tutoring in real-time simulations, explicitly as a virtual coach or implicitly as mentoring given by an agent in the scenario. Results from this work will help guide subsequent research and investment in the use of tutors and mentors in real-time dynamic, simulation-based training. 2. Synthetic Teammates - The use of synthetic agents as scenario participants is an area that is gaining increasing acceptance. The proposed effort will further the state of practice in this area, providing proof-of-concept demonstration of synthetic agents in Phase I and yielding data establishing the training benefits of this approach in Phase II. 2. Communications Training - The use of synthetic speech interaction is growing but remains limited, brittle and costly. The proposed effort will demonstrate techniques that show promise in overcoming these limitations for purposes of communications training, providing demonstrations in Phase I and robust implementation of this approach in Phase II. 4. Automated Assessment - An important touchstone in extending individual simulations to comprehensive training is the ability to automatically generate measures of user performance. Beginning with a preliminary implementation in Phase I and continuing with a suite of measures in Phase II, we will generate findings that document promising approaches to automated assessment, including the capture of spoken communication measures.
* information listed above is at the time of submission.