Todd J. Callantine
This paper describes methods and applications of intent inference for future teams of air traffic controllers that include a strategic planning controller responsible for 'conditioning' the traffic flow. The Crew Activity Tracking System (CATS) provides a framework for developing intent-aware intelligent agents to support controller teams. A proof-of-concept system provides reminders to the planner and another controller in real time. This team-level system draws upon related efforts to apply intent inference to better understand and model the planner’s task. These efforts entail enhancing CATS with a model of perception of the traffic display, and using different model forms within the CATS framework. The paper describes, specifically, inferring the planner’s strategy using a Bayesian Network model, and inferring the planner’s immediate intent using a temporal Bayesian model. The paper relates these efforts, and the team-level reminder system, to other relevant research.