Communicating pilot goals to an intelligent cockpit aiding system

In modern aircraft, the human pilots are no longer the only actors that control the aircraft and its systems. Machines, such as the autopilot and flight management system, also play an active role in control. In fact, several recent accidents occurred due to goal conflicts between human and machines. To facilitate the coordination of these actors, a computational aid called the AgendaManager (AMgr) was developed. The AMgr, which operates in a part-task simulator environment, attempts to facilitate the management of goals the actors are trying to accomplish and the functions being performed to accomplish them. To provide accurate knowledge of pilot goals for the AMgr, a goal communication method (GCM) was developed. The embedded GCM recognizes explicit and/or implicit pilot goals and declares them to the AMgr. The paper presents the development, architecture, operation, and evaluation of the GCM.