Hybrid Estimation of Distribution Algorithm Based Neuro-Fuzzy Dynamic Characteristic Modeling and Adaptive Control for Hypersonic Vehicle

摘要

The study on the modeling and high performance control for hypersonic vehicle has received great interest from the international research field of spacecraft. There are some alternative and complementary methods used to achieve this object. However, there are also some drawbacks in dealing with complex nonlinear cases. Here, a novel neuro-fuzzy dynamic characteristic modeling (NFDCM) method is designed and analyzed for the nonlinear longitudinal dynamics of a generic hypersonic vehicle. It combines the characteristic modeling method, which is an effective engineering-oriented modeling approach, with T-S fuzzy modeling method. Moreover, the proposed method also introduces the lowlevel learning power of neural network into the fuzzy logic system. Meanwhile, in the model, with the help of a novel Estimation of Distribution Algorithm based hybrid approach UMDA*, the key parameters identification and optimization are implemented. Based on the novel modeling method, the fuzzy adaptive controller for hypersonic vehicle is discussed. The proposed approach has been shown to be eective via simulation of velocity tracking task for vehicle.