论文部分内容阅读
讨论了载体位置无控、姿态受控情况下,空间机械臂姿态、末端爪手协调运动的自适应神经网络控制问题。由拉格朗日方法,建立了漂浮基空间机械臂的系统动力学方程。以此为基础,借助于RBF神经网络技术、GL矩阵及其乘积算子定义,对空间机械臂系统进行了神经网络建模。最后针对空间机械臂所有惯性参数未知的情况下,设计了空间机械臂载体姿态与机械臂末端爪手协调运动的自适应神经网络控制方案。提出的控制方案不要求系统动力学方程具有关于惯性参数的线性性质,且无需预知系统惯性参数的任何信息;并无需对神经网络进行离线训练、学习,因此更便于实时应用。一个平面两杆漂浮基空间机械臂系统的数值仿真,证实了该控制方案的有效性。
The control problem of adaptive neural network in which the attitude of the space manipulator and the coordinated movement of the end claw hand are discussed under the condition of uncontrolled position and controlled attitude are discussed. By Lagrange method, the system dynamic equations of the floating base space manipulator are established. Based on this, with the aid of RBF neural network technology, the definition of GL matrix and its product operator, neural network modeling of space manipulator system was carried out. Finally, aiming at the problem that all inertial parameters of the space manipulator are unknown, an adaptive neural network control scheme is designed to coordinate the attitude of the space manipulator and the hand claw of the manipulator. The proposed control scheme does not require that the system dynamics equations have linear properties about the inertial parameters and does not need to predict any information about the inertial parameters of the system. And it does not need to train and learn the neural network offline, so it is more convenient for real-time applications. The numerical simulation of a planar two-rod floating base-space manipulator system confirms the effectiveness of this control scheme.