Applications of Neural Networks in Aerospace Engineering
Summer Internship, 2015
Guide : Prof. A.K. Ghosh, Department of Aerospace Engineering, IIT Kanpur
We explored the various applications of Neural Networks in diverse fields of Aerospace Engineering. The project started with using artificial neural network training for parameter estimations in Hansa-3 aircraft.Further we explored the use of artificial neural network training in control problems wherein the elevator deflections are modelled using a 6-DOF system coupled with a trained artificial neural network.
​
Further we trained an artificial neural network for artillery shell trajectory prediction with the objective of improving the parametric prediction method that is presently being used.
An artificial neural network was trained to estimate the target of a Point Mass Model of an artillery shell for a given set of environmental conditions. The results are generated for a fixed bearing angle and varying bearing angles for estimating the target position and also estimating the required elevation angle and bearing for a given target position .The Point Mass Model was further improved to generate a Modified Point Mass Model which took into account additional factors such as spin induced drifts, coriolis forces of earth and better drag modelling. Neural network was trained to account for these changes and the resulting system was capable of predicting trajectory and target of an artillery shell with better accuracy than the parametric system that is currently being used.
​
Access the manuscript here.