Adaptive Neural Network-Based Fault-Tolerant Control for Quadrotor-Slung-Load System Under Marine Scene

Aiming at applying unmanned aerial vehicle (UAV) emergency transportation in marine scene, this paper investigates an adaptive neural network fault-tolerant control strategy for quadrotor-slung-load system (QSLS), in which actuator faults, marine wind disturbance and suspended payload are considered simultaneously. Firstly, dynamics model of QSLS and detailed marine wind field are established. Then, a fault-tolerant controller is designed by combining disturbance observer (DO) and radial basis function (RBF) neural network, in which external disturbances as well as lumped internal disturbances including system uncertainties and actuator faults are effectively inhibited. In addition, gradient descent algorithm (GDA) is introduced to train RBF to better approximate unknown dynamics and compensate the system. And Lyapunov stability analysis demonstrates convergence of the proposed strategy. Finally, the effectiveness and superiority of the proposed scheme are fully verified by comparative simulation and experimental tests.