A study of longitudinal control problems at low and negative damping and stability, with emphasis on effects of motion cues

As part of a general investigation to determine the effects of simulator motions on pilot opinion and task performance over a wide range of vehicle longitudinal dynamics, a cooperative NASA-AMAL program was conducted on the centrifuge at Johnsville, Pennsylvania. The test parameters and measurements for this program duplicated those of earlier studies made at Ames Research Center with a variable-stability airplane and with a pitch-roll chair flight simulator. Particular emphasis was placed on the minimum basic damping and stability the pilots would accept and on the minimum dynamics they considered controllable in the event of stability-augmentation system failure. Results of the centrifuge-simulator program indicated that small positive damping was required by the pilots over most of the frequency range covered for configurations rated acceptable for emergency conditions only (e.g., failure of a pitch damper). It was shown that the pilot's tolerance for unstable dynamics was dependent primarily on the value of damping. For configurations rated acceptable for emergency operation only, the allowable instability and damping corresponded to a divergence time to double amplitude of about 1 second. Comparisons were made of centrifuge, pitch-chair and fixed-cockpit simulator tests with flight tests. Pilot ratings indicated that the effects of incomplete or spurious motion cues provided by these three modes of simulation were important only for high-frequency, lightly damped dynamics or unstable, moderately damped dynamics. The pitch- chair simulation, which provided accurate angular-acceleration cues to the pilot, compared most favorably with flight. For the centrifuge simulation, which furnished accurate normal accelerations but spurious pitching and longitudinal accelerations, there was a deterioration of pilots' opinion relative to flight results. Results of simulator studies with an analog pilot replacing the human pilot illustrated the adaptive capability of human pilots in coping with the wide range of vehicle dynamics and the control problems covered in this study. It was shown that pilot-response characteristics, deduced by the analog-pilot method, could be related to pilot opinion. Possible application of these results for predicting flight-control problems was illustrated by means of an example control-problem analysis. The results of a brief evaluation of a pencil-type side-arm controller in the centrifuge showed a considerable improvement in the pilots' ability to cope with high-frequency, low-damping dynamics, compared to results obtained with the center stick. This improvement with the pencil controller was attributed primarily to a marked reduction in the adverse effects of large and exaggerated pitching and longitudinal accelerations on pilot control precision.