Intelligent Control Method for Hypersonic Vehicle Considering Time Frequency Domain Performance Indicators
Aiming at the attitude control problem of hypersonic vehicle with wide flight envelope and strong aerodynamic uncertainty,an intelligent optimal control method considering time frequency domain performance Indicators is proposed.This method includes two parts:offline design and online design.In the off-line process,according to hypersonic vehicle dynamics model and linear quadratic regulator(LQR)control design method,a large number of attitude control simulation sample data sets are generated,the control performance index evaluation model is constructed,the optimal control parameters are screened,and the neural network mapping relationship with flight state and aerodynamic uncertainty as inputs and optimal LQR control parameters as outputs is established.In the online process,the aerodynamic uncertainty of hypersonic vehicle is captured through real-time aerodynamic parameter identification,and the intelligent parameter optimization network is used to make the attitude control of hypersonic vehicle still the optimal control parameter to meet the control index under the condition of aerodynamic uncertainty.Finally,simulation results show that the proposed method can effectively improve the robustness of hypersonic vehicle attitude control under wide flight envelope and strong uncertainty.
Hypersonic vehicleAttitude controlDeep neural networksOptimum control
NETL
NSTL
万方数据
