The Reconfiguration Of The Self-Repairing Flight Control Law For Fighters Using Neural Networks

The objective of this thesis is to design an artificial neural network adaptive reconfigurable flight control law basing on the view of engineering and to establish a flight simulation system with normal aircraft, fault aircraft, normal flight control law, reconfigurable flight control law and fault verification by C++ program.First, to research the reconfigurable flight control system, a flight simulation system of normal aircraft with normal flight control law is established. Using artificial neural network, a new approach to process flight dynamic datum is presented, and a fault aircraft flight simulation system is established.Second, the thesis describes an adaptive reconfiguration method for augmenting an exist linear controller by an artificial neural network. The key feature of the proposal approach is using feedback linearization method to design a single hidden layer artificial neural network whose weights is renewed online to augment parametric uncertainty and unmodeled dynamics. There is a good future in real engineering applying, because there is no need to redesign a new "inverse controller".Third, using augment information that product by reconfigurable control law, an artificial neural network fault verification method is presented.Finally, Resource codes of the flight simulation system, which are 11,000 lines, are established. Results of simulation test have demonstrated that the method in this thesis is suit to not only the normal aircraft, but also the fault aircraft. Self-Repairing Flight Simulation Platform verifies the flight simulation system.