NN adaptive backstepping based aircraft reconfigurable control design

Abstract

Reconfigurable flight control could maintain the safety and performance of aircraft under failure or damage situation. In this paper, aircraft reconfigurable control is designed using NN based adaptive backstepping method. Backstepping is used to build the baseline control law. The model changes due to failures or damage are considered as modeling errors and identified online using B-spline neural networks. Closed-loop stability is guaranteed through the backstepping design and update laws of the network weights. The requirement for aircraft model under which the reconfiguration is realizable is given. Based on this, the feasibility of control reconfiguration is analyzed. Simulation results show the effectiveness of the control law.