空间机器人的增广自适应神经网络补偿控制.pdf

第 23 卷第 12 期 系统仿真学报© Vol. 23 No. 12 2011 年 12 月 Journal of System Simulation Dec., 2011 空间机器人的增广自适应神经网络补偿控制 陈志勇, 陈 力 (福州大学机械工程及自动化学院，福州 350108) 摘 要：研究了具有外部扰动空间机器人系统的补偿控制问题。首先，给出了漂浮基空间机器人系 统的动力学方程。进而借助于增广变量思想，针对系统存在有未知惯性参数及外部扰动的复杂情况， 先后设计了空间机器人系统关节运动、末端爪手运动的增广自适应神经网络补偿控制方案。所提控 制方案无需对载体的位置、线性速度和线性加速度进行实时地测量与反馈；且较比传统自适应控制 方法，又不要求动力学方程满足关于系统惯性参数的线性函数关系，因此有效避免了对系统回归矩 阵进行繁琐提取的麻烦。仿真运算，验证了上述控制方法的有效性。 关键词：空间机器人；外部扰动；关节运动；末端爪手；增广变量思想；神经网络补偿控制 中图分类号：TP241 文献标识码：A 文章编号：1004-731X (2011) 12-2750-06 Augmented Adaptive Neural Network Compensation Control of Space-based Robot CHEN Zhi-yong, CHEN Li (College of Mechanical Engineering and Automation, Fuzhou University, Fuzhou 350108, China) Abstract: The compensation control problems of space-based robot system with external disturbances were studied. The dynamic equations of the system were given firstly. Then, with the augmentation approach, the augmented adaptive neural network compensation control schemes were developed for space-based robot with unknown inertial parameters and external disturbances to track the desired trajectories in joint space and in inertial space, respectively. The control schemes proposed don’t require the measurements and feedback of the position, linear velocity and acceleration of the base. Besides, compared with traditional adaptive control schemes, they can avoid linearly parameterizing the dynamic equations of the system and subsequently reduce tedious computations during the determination of the regression matrix for the system. The simulation