论文部分内容阅读
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(CNN)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。其次,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(NFTSM)控制器。为提高控制精度,引入了只需要期望信号的CNN来估计系统总扰动,从而有效削弱了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(ANN)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。
Aiming at the problem of dynamic non-edge trajectory imaging of the point of interest along the ground under the stars, a non-singular fast terminal sliding mode controller based on the Chebyshev neural network (CNN) was designed. Firstly, the attitude adjustment method for non-orbit imaging modes is studied and the corresponding desired attitude angles and attitude angular velocities are derived. Second, a non-singular fast terminal sliding mode (NFTSM) controller is designed based on the tracking kinematics model described by the quaternion of errors. In order to improve the control accuracy, CNN, which only needs the desired signal, is introduced to estimate the total system disturbance, which effectively weakens the inherent chattering of the sliding mode system. In order to ensure the output of neural network is bounded, a switch function is introduced to realize the switching control between adaptive neural network (ANN) and robust control. Finally, the numerical simulation of the attitude control system with disturbance and parameter uncertainty shows that the proposed method has the advantages of fast convergence and high control precision, which has certain practical significance.