Assessment of Manual Flying Skills by Combining Aircraft Parameters with Pilot Control Inputs

Abstract

In a recent full flight simulator experiment, 57 commercial airline pilots (long-haul captains and short-haul first officers) had to accomplish a realistic scenario ending in a manual flown “raw data approach” (localizer and glideslope capturing with absence of autopilot and flight director aid). The manual flight performance was evaluated by an instructor. Various flight path parameters and pilot control input data were collected from the simulator for an objective evaluation and comparison with the instructor assessment. At first the outer control loop was analyzed by using flight path tracking errors. This showed significant difference between the two groups of pilots. Additionally the pilot steering (roll and pitch) strategy and the aircraft reaction were analyzed in the frequency domain to analyze the inner control loop. The steering strategy for pitch consist significant lower frequency than for bank in both groups. The results showed that pilots used steering inputs with higher frequency than the aircraft reaction frequency. Pilots with lower instructor evaluation show significant higher portions in this ineffective frequency band. The results show that a combination of outer and inner loop parameters is a good indicator to evaluate pilot manual flying skills.