Adaptive Flight Controller Design Based on RBF Network

This paper designs a self-adaptive controller for the short-period longitudinal mode of an aircraft. At first, the parameter space method is used to find admissible set of feedback controller parameters for 40 different flight conditions in the flight scope. Then, all designed parameters are used for training a radial basis function (RBF) network. In this way, the network can remember and output the designed parameters for relevant velocity and altitude. Thus, the controller parameters can be regulated adaptively by changing flight conditions. Simulation results show that the control system meets the military specifications for all 40-flight conditions. At the same time, it indicates that the technique proposed in the paper is effective and essential