Actuator fault tolerant controlling using adaptive radical basis function neural network SMC for quadrotor UAV

摘要

In this paper, an actuator fault tolerant controller for the quadrotor UAV is designed. The new method is based on the adaptive RBF neural network and sliding mode control. The dynamic equations which includes external disturbances are constructed by Netwon-Euler theorem. External disturbances of unknown upper bound are approximated by RBF neural network. The adaptive strategy realized the on-line estimation of the actuator fault. The integrated FTC design approach guarantee the state variables converge to the desired values at limited time. The Lyapunov function prove that the system is global asymptotically stable. Futherfore, the backstepping sliding mode control technique alleviates the chattering phenomenon which result from switching law and solve the "explosion of complexity". The simulation results testified the effectiveness and robustness of the presented methodology for the quadrotor UAV.