关键词: 动态面控制, 鲁棒控制, 近空间飞行器, 神经网络

Abstract:

Considering the severe changes of aero － dynamic parameters and susceptibility on the external disturbances of nearspace vehicle
( NSV) ，a backstepping control strategy with robust adaptive dynamic surface is proposed in this paper for the disturbance problem in the NSV
longitudinal trajectory system during the hypersonic process． Firstly，the complex nonlinear longitudinal dynamic model is transformed into
nonlinear non － affine model by using the input － output feedback linearization method． Then，the“computer explosion“in the computation
procedures for derivatives is avoided with the estimation of the virtual control law in one － order low pass filter． The robustness item in the virtual
controller combined with approximation capability of neural network is used to eliminate the parameter uncertainties and external disturbances
in NSV． The simulation results show the improved robustness in this method with reduced complexity．

Key words: dynamic surface control, Robust control, nearspace vehicle, neural network