On-line Identification of longitudinal Aircraft Dynamics Using Extended Neural Networks

Document Type : Original Article


1 Eng. Amgad M. Bayoumy, Military Technical College (MTC), Cairo, Egypt.

2 Assoc.Prof. Fawzy Ibrahim AbdelGhany, Chairman of Radar and Guidance Department, MTC, Cairo, Egypt.

3 Prof. Ibrahim Mansour Ibrahim, Chairman of Design Center, Ministry of Military Production, Cairo, Egypt Military Technical College Kobry Elkubbah, Cairo, Egypt.


In this paper Artificial Neural Networks (ANN) approach is used to identify the longitudinal aircraft dynamics online. Memory Neuron NN (MNN) is used here as a model to perform online non-parametric identification of aircraft dynamics. The aircraft longitudinal motion is modeled as a linear system. The input and output of the aircraft model are used as training pair to the NN model. NN is extended by adding two more parameters, the maximum input and maximum output vectors, to learn the changes in I/O bounds dynamically. After the training phase the NN output shows good agreement with the aircraft output for the same input.