AN INTELLIGENT CONTROL PARADIGM FOR FLIGHT CONTROL APPLICATIONS

摘要

A control approach was developed for nonlinear systems based on aircraft modeling, and control performance is based on the accuracy of the model. In the actual aircraft implementation, a neural network is substituted for the model and trained offline to produce the same results. A methodology is developed to validate both the network and the overall controller as meeting performance and reliability requirements. Development of a multi-axis controller using this methodology with a fuzzy controller for reconfiguration is presented. Computer simulations to test the methodology on a simple airfoil and a pitch axis flight controller are also reported.