06.1 - Flight Dynamics and Control (Control & Modelling)
SYSTEM IDENTIFICATION USING FLIGHT TEST DATA AND MACHINE LEARNING
R Larsson, Saab Aeronautics, Sweden
The usability of flight mechanical simulations relies on sub-models. A method for estimating nonlinear aerodynamic properties, using flight test data in combination with machine learning is investigated. Here, a subscale radio-controlled aircraft is used. Neural Network functions for system identification are possible. It can handle nonlinear system behavior but relies heavily on data available.