Pilot performance in automated cockpits: a comparison of moving and nonmoving thrust levers

Abstract

Two different design-philosophies with respect to the Man-Machine Interface of current auto thrust systems were compared in a flight simulator study. With the so-called moving thrust lever concept, the positions of the thrust levers are servo-controlled by the auto thrust system. Hence, the thrust levers provide the pilot with additional information regarding actions taken by the auto thrust system. In contrast, the so-called nonmoving thrust levers are not servo-controlled by the auto thrust system. Pilot performance was investigated when operating these thrust lever configurations. Eight recently licensed civil pilots flew both thrust lever configurations in a moving base flight simulator. The variables used to describe pilot performance were: number of correct detections of problems and failures, detection time and thrust awareness. Questionnaires were used to measure the preference of the pilots for both auto thrust systems. The study demonstrates that moving thrust levers provide effective feedback and assist the pilot in anticipating certain malfunctions and deviations. Pilot detection times are typically shorter and pilots are more certain about their thrust estimates. Pilots confronted with both systems for the first time prefer the moving thrust levers after flying with it. Surprisingly, lack of motion of the moving thrust levers was a strong stimulus for pilots when they expected such a motion to occur. This could indicate that the thrust levers served the pilots especially by providing them prompt feedback on intended actions and served less as a device alerting the pilot to unexpected events.