Abstract: In order to resolve the problem of wind disturbance and model uncertainties during flight control of stratospheric airships, a nonlinear robust trajectory control method with combination of backstepping and neural network is proposed. The Lagrange dynamic models for stratospheric airship are established, the control laws for thrust and torque are directly obtained by using backstepping, and the optimal control gain parameters are dynamically updated with the RBF neural network. The numerical simulations show that the new control method can make the stratospheric airship accurately track the combined linear/circular reference trajectory, and overcome the adverse effects caused by wind disturbance and model uncertainties.

Key words: Stratospheric airship, Trajectory control; Backstepping, RBF neural network