反步法与神经网络融合的平流层飞艇轨迹鲁棒控制方法

Robust Trajectory Control Method for Stratospheric Airships with Combination of Backstepping and Neural Network

针对平流层飞艇水平轨迹控制面临外部风场扰动和模型参数不确定的特点,提出一种反步法与RBF神经网络融合的非线性鲁棒轨迹控制方法。该方法采用Lagrange动力学模型,通过反步法直接求解推力和力矩的控制律,采用RBF神经网络动态优化调整反步法的控制增益参数。仿真结果表明,新的控制方法可实现平流层飞艇对直线/圆组合式参考轨迹的高精度跟踪控制,有效克服风场扰动和模型不确定性因素造成的不利影响。

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.