Neuro-controller for satellite attitude control

number: 
590
English
Degree: 
Author: 
Omar Muwaffaq Mustaff
Supervisor: 
Dr. Mohammed Z. Al-Faiz
year: 
2001
Abstract:

The attitude dynamics of a dual-spin stabilized satellite in a small angles maneuvering mode which represents a multivariable coupled dynamics, was introduced and a neuro-control strategy was built for the attitude control problem. The proposed neurocontroller makes use of the desired attitude, feedback error, actual attitude and their appropriate time delays to generate the required control action. A very fast training algorithm, namely, Levenberg-Markquardet algorithm was used to enhance the activity of the proposed controller and hence the problems that may occur due to the use of the backpropagation algorithms were avoided. From the computer simulation results, which were carried out using MATLAB 6, a good tracking performance, coupling cancellation and a fast attitude mutation motion damping were obtained. Also, a good robustness performance against attitude dynamic parameters variation is achieved. In addition, the proposed controller is successfully used with a three-axis stabilized satellite model.