变结构航天器模糊神经网络滑模控制器设计

变结构航天器是目前航天领域的重要发展方向，航天器结构的变化将导致质量分布发生明显变化，这对航天器动力学建模和控制器设计都提出新的问题。针对这种情况，采用混合坐标法和拉格朗日方程建立了航天器刚柔耦合动力学模型，利用几种典型工况的参数近似得到变结构过程中动力学参数的变化规律。设计滑模控制器对航天器变结构过程进行姿态控制，为提高滑模控制器的适应性，设计模糊神经网络（FNN）自适应调节滑模控制器参数，并利用径向基函数（RBF）神经网络逼近动力学模型，得到控制力矩与姿态变化之间的近似关系，用于FNN的优化。通过仿真得到航天器变结构期间无控、滑模控制和模糊神经网络滑模控制的姿态变化，仿真结果对比验证了模糊神经网络滑模控制对于滑模控制的优势，证明了其在变结构航天器姿态控制方面的有效性。

Sliding mode control based on fuzzy neural network for variable structure spacecraft

Variable structure spacecraft is an important development direction of astronautics. The mass distribution of the spacecraft will change significantly during the configuration variation, and this will generate new problems of dynamic modeling and controller design. To solve these problems, the hybrid coordinate method and Lagrange equation were used to build the dynamic model of the spacecraft, and changing rule of the dynamical parameters was approximated by several typical working conditions. Sliding mode controller was designed to control the attitude during the variation of the spacecraft. To improve the adjustment of the controller, fuzzy neural network(FNN) was designed to adjust the parameters of the controller adaptively. The radial-based function (RBF) neural network was designed to approximate dynamic model, and thus the relationship between the control torque and attitude variation was obtained, which was used to optimize the FNN. The attitude of the variable structure spacecraft during the structure variation with no control, sliding mode control and fuzzy neural network was acquired in simulation. The results verify the effectiveness of the fuzzy neural network adaptive sliding mode controller, and comparisons were made to verify the good properties of the proposed controller.