Intelligent position control of earth station antennas with model independent friction compensation based on MLP neural networks

An Intelligent Control (IC) method based on MLP neural networks and Through Model Lemma (T.M.L.) is implemented to control the position of a Low Earth Orbit (LEO) satellites tracking earth station antenna. This approach relies on two different multilayer neural networks with delayed inputs, for the purpose of identification and control. Nonlinear term in motors caused by friction is not necessary to be measured or identified for considering in the controller designing using this method. However due to test of the proposed method performance, this nonlinearity term modeled by a dead zone block. Simulation results show the effectiveness of T.M.L. method for robust control in the presence of friction nonlinearity.