摘 要：针对地面兴趣点不沿星下点轨迹的动态非沿轨迹成像问题，设计了一种基于切比雪夫神经网络的非奇异快速终端滑模控制器。首先研究了非沿轨迹成像模式的姿态调整方法，并推导了相应的期望姿态角和姿态角速度。基于由误差四元数描述的跟踪运动学模型设计了非奇异快速终端滑模控制律(NFTSMC)。进一步为提高控制精度，引入只需要期望信号的切比雪夫神经网络(CNN)来估计系统总扰动，从而有效削弱滑模系统的固有抖振。为保证神经网络的输出有界，引入一个开关函数以实现自适应神经网络与鲁棒控制之间的切换控制。最后对具有干扰和参数不确定的姿态控制系统进行了数值仿真，结果表明该方法收敛速度快，控制精度高，具有一定的工程实际意义。

Abstract: A nonsingular and fast terminal sliding mode controller based on the Chebyshev neural network (CNN) is designed for the next-generation smart imaging satellite, whose ground targets and ground trace are not parallel. First, the specific method of attitude adjustment for smart imaging mode is studied to get the desired attitude angle and angular velocity. According to the t racking error dynamics and kinematics described by unit quaternion error, a nonsingular and fast terminal sliding mode controller is derived. Considering the disturbance rejection, a Cheby-shev neural network whose basis functions are implemented using only desired signals is introduced to approximate the general disturbance which efficiently weaken the chattering inherent in sliding mode structure. In order to guarantee that the output of the NN used in the controller is bounded by the corresponding bound of the approximated disturbance, a switch function is applied to generate a switching between the adaptive neural network control and the robust controller. Finally, numerical simulations on the attitude tracking control of spacecraft in the presence of environmental disturbance and parameters uncertainties are performed, whose results show the designed control scheme can meet the requirements of control precision for the smart imaging with fast convergence speed and good robustness. The system can stabilize within and the precision of attitude angle and angular velocity is and ，respectively.