Document Type


Publication Date





The interest in determining pilot’s workload has increased, mainly when considering the human factors certification issues of new aircraft, where the high insertion of automation devices has been proposed to increase the mental workload (MW). The use of subjective and physiological measurements for pilot workload evaluation has been previously investigated in a simulator, and further tested during real flights, but no comparison has been performed between the results of such different situations. This study compared the data of subjective and physiological measurements obtained on the evaluation of MW in simulators, with those measured during real flights. Seven pilots performed six different tasks in the simulator, and two pilots flew six flights. These flights were performed aiming at the certification process of a new aircraft regarding human factors, and, as certification requirements, they were conducted with the aircraft set under abnormal condition. In the simulator all pilots performed the same tasks, while assuming both, the pilot flying (PF) and pilot monitoring (PM) positions. During the flights the pilots exchanged the PF and PM position depending on flight scheduled. The pilots’ ECG was registered in a computer, and the heart rate variability (HRV) processed for each task of simulator, and take-off and landing of flights. From the power spectral density function of HRV it was determined the total power of low frequency (LF) and high frequency (HF). The HRV was analyzed as the ratio of the LF/HF obtained during the evaluated task (tasks of simulator; take-off and landing of flights) and the LF/HF values obtained during a rest test, which was registered either prior to the flight or to the section of simulator. The NASA-TLX Scale was applied just after the task was finished. The results showed the NASA-TLX score of PM in the simulator to give higher values than PF, but this was not statistically significant. The same tendency was observed on HRV, but most of the tasks reached statistical significance (p<0.05). Conversely, during the flights PF showed higher values of NASA-TLX score and HRV than PM, although no statistical test was applied in such data due to the number of pilots. Surprisingly, the NASA-TLX showed higher values in simulator than during the flights. The data from simulator tasks and flights suggests the mental demand of NASA-TLX to be correlated to HRV, but HRV appears to be more sensitive to different intensities of mental workload. The differences observed when comparing simulator and flights were more likely due to the nature of the tasks, level of automation, or even the workload perceived by pilots. It is concluded that the use of both subjective and HRV measurements give a potential tool in the evaluation of pilot MW, and could further be used to quantify the workload objectively, defining acceptable ranges of MW that pilots are subjected to. The assessment could be performed in simulator since this machine presented results very compatibles with those found in real flights.