Application of Neural Networks for Flight Simulation

摘要

Application of neural networks for direct prediction of lateral-directional force and moments coefficients from the measured flight data of the research aircraft is proposed in this paper. Proposed model of neural networks appears to be a suitable practical approach to develop relationship between flight variables. This relationship eliminates the need of aerodynamic model as well as thrust model to develop accurate flight simulation software for the aircraft. The validation of proposed feed forward neural networks based technique is done with simulated as well as measured flight data of the aircraft.