基于自适应RBF神经网络的可重复使用运载器再入段姿态控制

Research on Attitude Control of Reentry for Reusable Launch Vehicles Based on Adaptive RBF Neural Network

针对再入段可重复使用运载器(Reusable Launch Vehicles,RLV)姿态控制问题,提出一种基于自适应径向基函数(Radial Basis Function,简称RBF)神经网络(ARBFNN)的姿态控制方法。首先,建立RLV的6-DOF非线性动态模型,并将旋转动力学模型变成严反馈形式。然后,设计了一种自适应RBF神经网络控制(ARBFNN)结构,可以减少神经网络逼近误差的影响。同时,通过李雅普诺夫和自适应控制的结合消除了控制器设计过程中的不确定影响,并验证了系统的稳定性。最后,通过仿真验证了所提算法在解决再入段RLV姿态控制问题上的有效性。

An adaptive radial basis function neural network control （ARBFNNC） method is presented to address the attitude control problem of reentry for reusable launch vehicles （RLV）. Firstly, the 6 - DOF nonlinear dynamic model of RLV is established and the rotational dynamics and rotational kinematic are transformed into strict -feedback system. Hence, an ARBFNNC system structure is established which can benefit to reduce the effects of the approximation error of neural network. Then, by combining the adaptive method with Lyapunov function, the stability of system is guaranteed and the uncertain effect in the process of controller design is eliminated. Finally, a numerical simulation is employed to illustrate the effectiveness of the proposed strategy by solving problem of attitude control accuracy and the system robustness. The simulation results demonstrate that the ARBFNNC system can achieve favorable control performance after the parameter learning of the ARBFNN.