摘要
对于双臂 n自由度(n为正整数)空间机器人的姿态控制问题,在控制算法中,利用再生核替代传统的Fourier基函数,能克服 Fourier变换不能用于局部分析的缺点,且对奇性敏感,求解稳定,并能提高控制精度,因此该文给出了一种基于再生核理论的最优控制算法.同时给出了双臂n自由度空间机器人广义雅可比矩阵的求导公式.该求导公式较以往的推导方法具有显式、易求的特点.在数值试验中,针对双臂三自由度空间机器人的模型进行了仿真,结果表明,利用文中提出的再生核解法解决双臂空间机器人的姿态控制问题是有效的.
The problem of attitude control for the dual-arm space robot is discussed in this paper. The reproducing kernel is here used to substitute for the classical Fourier basic function. It overcomes the shortcoming that the Fourier transformation can not be used in the local analysis. It is sensitive to singularity and has stable error in solution. The proposed method can improve the precision of controlling. Therefore, an optimal control algorithm is presented based on the theory of reproducing kernel. The generalized Jacobian matrix (GJM) of the dual-arm space robot of the n-degree of freedom is deducted simultaneously, whose formula is more explicit and easier to solve than those deducted by other methods. After degree of freedom n is fixed, the parameters of the robot model can be directly put into the expression of GJM so as to get the value of GJM. It is a new research result that is worth popularizing. In numerical simulation, experiment about the optimal control of the dual-arm space robot of the three-degree of freedom is carried out. The result indicates that the reproducing kernel method is efficient to attitude control of the dual-arm space robot's body. Besides, it can attain higher controlling precision than traditional Fourier method, with a higher precision of over 10 times and it has advantages over the wavelet expansion approach, such as simpler operation, easy to be mastered and used.
出处
《计算机学报》
EI
CSCD
北大核心
2005年第2期232-240,共9页
Chinese Journal of Computers
基金
国家"八六三"高技术研究发展计划项目基金(863 2 4 1 2)资助.
