任务空间内不确定性空间机器人智能控制方法

Intelligent Control Method for Uncertain Space Manipulator in Task Space

针对自由漂浮空间机器人首先给出其任务空间内的动力学模型,并由此利用GL矩阵推导出了空间机器人任务空间内神经网络模型。采用了只含有空间机器人系统尺寸信息的雅可比矩阵来近似代替广义雅可比矩阵,将此带来的影响视为一种不确定性的影响,设计了一种神经网络自适应控制律,利用神经网络去逼近整个系统的不确定项和系统的动力学模型,解决了任务空间内无需系统名义动力学模型和不确定性等先验知识的控制问题,并给出稳定性的证明。最后给出了两自由度空间机器人的仿真结果,其证实了该方法的有效性。

The free-floating space manipulator dynamic equations and neural network model in task space based on Graphic Language（ GL） matrix have been proposed. Neural network and adaptive control were combined to design a control law for uncertain space manipulator dynamic equation,with a usual Jacobian matrix of size parameters to approximate the generalized Jacobian matrix. The control problem without the system parameters and uncertain model was solved,and the proof of stability was given. Finally,the simulation results have shown that the proposed method is validated for two dimensional（2D） free-floating space manipulator.