06.1 - Flight Dynamics and Control (Control & Modelling)
FAULT TOLERANT CONTROL SYSTEM OF AIRCRAFT BY FEEDBACK ERROR LEARNING WITH SYSTEM ESTIMATION
N. Yoshikawa, The university of Tokyo, Japan
A neural network-based fault tolerant control system of an aircraft assuming aileron failures was investigated.rnThis system is divided into 2 neural network (NN) blocks, a fault detector and adaptive controller.rnThe fault detector is constructed with recurrent neural network.rnThe adaptive controller is learned by the reinforcement learning method.