This paper presents a neural network based controller used in commanding time varying systems with uncertainties task First, a reduction procedure of the initial set of parameters using an unsupervised pattern recognition technique was applied. After this a feed-forward neural network was trained using the minimized set of data. The advantage of this method is over-passing of the difficulties implied by the direct solving of the differential models, which are necessary in a classical approach. An application of a missile-target tracking was implemented using the mentioned method, and the results are compared with those obtain in a classical approach.
展开▼
