New Approach for the Identification and Validation of a Nonlinear F/A-18 Model by Use of Neural Networks