Abstract
The purpose of this study is to analyze human pilot behavior at landing operation and compare the results. The pilot model is constructed using neural network, and is trained with the results of the simulator tests. The structure of neural network is determined by GA. The inputs to this model are observation cues such as the horizon and the runway, and pilot control inputs. The outputs are the control column angle and the engine throttle angle. Finally, the importance of each input is evaluated using Contribution Ratio Analysis and Sensitivity Analysis from the obtained neural network models. We analyzed the data of veteran pilots and newcomer pilots, and compared the result.
