A theory of performance measurement for operator controlled systems is presented. The theory permits synthesis of a system performance measure which scores performance on successive data samples based on the impact of the sampled performance on the overall summary of performance. Since performance is measured and evaluated on each sample, the dynamics of the controlled element, i.e., the aircraft, are effectively removed from the measurement eventhough the pilot (operator) continues to control the aircraft. While the theory directly applies to problems where the performance limiting factors are known, the method has been extended to apply to problems where the performance limiting factors are not known explicitly, but are known to be implicit in the performance data. This paper documents the development of measures for aircraft carrier landings for the glide path and angle of attack control channels. Flight data obtained from the Visual Technology Research Center, Naval Training Equipment Center, Orlando, Florida, was analyzed using the measures. The data on carrier landings were available on 9-track magnetic tape consisting of flights by four subjects each performing on 42 flights. Each flight was performed on a particular combination of glide path error display and day/night combination. Two types of glide path displays were used, resulting in four treatments (two displays and two light conditions). One display was the conventional glide path display and the other was a command display which incorporates error rate information with glide path error presentation. Each subject controlled the aircraft to a carrier landing three times with each treatment. The resulting performance scores were aggregated by subject, treatment, range to carrier deck, and error and error rate cells. The results suggest that the command display offers improved glide path control especially during day light conditions. Further, the measurement technique was shown to permit determination of the relationship between performance and factors that change during the flight - such as error and error rate magnitudes. The measurement technique bypasses the time delays due to aircraft dynamics - by accounting for them - permitting evaluation of performance on short (less than a second) intervals of flight. Consequently, observed differences in performance at different error and error rate magnitudes were detected suggesting that significant non-linear control techniques may have been exhibited by the pilots.
