基于高阶神经网络扩展卡尔曼滤波器逆算法的非线性挠性结构的姿态控制

本文针对非线性挠性结构的姿态控制，提出了一种基于高阶神经网络及径向基函数网络（ＲＢＦＮ）相结合的网络模型，用于非线性挠性结构的动态系统辨识，以及基于卡尔曼滤波器（ＥＫＦ）逆算法的控制策略。针对神经网络辨识时的模型误差，提出了一种简单有效的补偿方法，给出了建模误差补偿与补偿时仿真结果。仿真得出，该方法具有收敛快，算法简单，并能有效消除建模误差等优点。

Attitude Control of Nonlinear Flexible Structure Based on Inversion Algorithm of Extended Kalman Filter of High-Order Neural Networks

According to the attitude control of nonlinear flexible structure,a high-order RBF neural network is presentedin this paper which combines high-order neural network and Radial basis function network (RBFN),The high-order RBFN isused for dynamical identification of nonlinear flexible structure and for the design of control scheme based on inverse algorithmof extended Kalman filter(EKF),The modeling error can be xompensated by a simple but efficient method that avoids theoccurrence of divergence in the inverse algorithm,and the comparison of results is presented between modeling error compensatioo and no modeling error compensaton.According to the simulation,it is explicit that this method is of these characters suchas fast convergence and simple algorithm.