An experimental study on attitude control of spacecraft using neural networks

Abstract

This primary objective of this study is to demonstrate ground-based experiment for the attitude control of spacecraft. A two-axis rotational simulator with a flexible arm is constructed with on-off air thrusters and reaction wheel as actuation. The simulator is also equipped with payload pointing capability by simultaneous thruster and DC servo motor actuation. The experiment of spacecraft attitude control is performed using the reaction wheel. Using the reaction wheel the performance of the PID and the three proposed neural network controllers is illustrated by simulation and experimental results for a two-axis rotational simulator.