基于RBF网络上界自适应学习的预警卫星滑模控制

Early warning satellite sliding mode control based on RBF neural networks adaptive learning

分析了RBF(径向基函数)神经网络的基本结构和数学特性,对于预警卫星动力学系统的不确定性上界值无法测量和未知的情况,采用RBF神经网络可以对较强干扰上界进行自适应学习,并可降低控制和动力学带来的抖振。针对带有摆镜的预警卫星姿态控制问题,提出了一种基于神经网络扰动补偿的姿态滑模控制方法。针对RBF网络正交最小二乘(OLS)学习算法,采用RBF神经网络来学习不确定因素的上界值,并设计了预警卫星的姿态控制规律,解决了预警卫星动力学扰动补偿问题。利用数值仿真估算了基于RBF网络上界自适应学习滑模控制的预警卫星姿态控制系统的性能指标。

The basic structure and mathematics characteristic of RBF （Radius basic function） neural network are studied. In the conditions of the uncertain up bound value can not be measured properly and unknown for early warning satellite dynamic system, adopting RBF neural network adaptive learning the large disturbance up bound value, and reducing the vibration of control and dynamics. For the early warning satellite with motive mirror attitude control problem, a kind of attitude sliding mode control method based neural network disturbance compensation is provided. For RBF network orthogonal least square （OLS） learning algorithm, adopt RBF neural network to learning uncertain up bound value, and design attitude control law of ear!y warning satellite, and solve the dynamics compensation problem of early warning satellite. Mathematic simulation is used to estimate early warning satellite attitude control system guideline based RBF network up bound value adaptive learning sliding mode control.