A prototype airplane takeoff monitor with learning features

This thesis resulted in the development of a prototype instrument for airplane take-off performance monitoring with learning feature. The most important feature for this instrument is the learning of the real airplane take-off performance and using the historical data of this particular airplane to predict take-off performance for present conditions instead using a pure theoretical model. The method and program developed here can be directly use in a commercial type instrument. The system can run independently or be integrated into a Flight Expert System(FLES)(¹). The instrument automatically compares measured acceleration with that predicted using historical data. The flight crew's judgment for go will be objectively reconfirmed if the data agree well from the early stages of ground roll until rotation. This will help to insure no false rejected take-offs. On the other hand, if low acceleration is detected, a yellow or red warning will be issued immediately depending on the seriousness of the condition. The measured data together with flight crew's experiences and quick judgment will insure a cautious take-off or rejected take-off at lower speed. After the takeoff, this new data may be added to the data base for future use.