An Automated Simulation Pilot Capability to Support Advanced Air Traffic Controller Training

Abstract

High Air Traffic Controller attrition and the transition to NextGen (Next Generation Air Transportation System) necessitate the use of innovative methods to improve the delivery and quality of ATC training while reducing costs. The MITRE Corporation’s Center for Advanced Aviation Systems Development (CAASD) and the Federal Aviation Administration (FAA) have been conducting research over the past several years to determine where improvements can be made in the Air Traffic Control (ATC) training process. To support this research, CAASD developed a high-fidelity ATC training prototype (known as the enrouteTrainer) and conducted training on Developmental Controllers at the FAA’s Indianapolis Air Route Traffic Control Center (ARTCC). Initial assessments confirm the anticipated benefits. The ATC training prototype includes an autonomous simulation-pilot capability that uses automated speech recognition and speech synthesis to mimic the role of human sim-pilots in traditional ATC training systems. CAASD defined performance measures for evaluating the realism of the autonomous simulation-pilot, with respect to speech recognition accuracy and appropriate simulated pilot behavior. During the field evaluations, several iterations of evaluation and refinement were made to the speech recognition algorithms to improve speech recognition accuracy. This paper describes the speech recognition performance measures and presents results collected during field evaluations conducted at the Indianapolis ARTCC in June 2007.