空间机器人抓捕目标后动力学参数辨识研究

Dynamic Parameter Identification of Free-Floating Space Robots after Capturing Targets

针对自由漂浮空间机器人抓捕目标后的动力学参数辨识问题,提出一种参数辨识的持续激励轨迹设计方法。首先,基于动量守恒原理建立了自由漂浮空间机器人的动力学参数辨识模型;然后,采用有限傅里叶级数对空间机器人的机械臂关节运动轨迹进行参数化表示,并以参数辨识回归矩阵条件数最小化为指标,通过求解一个包含多约束的非线性优化问题得到傅里叶级数的待定系数;最后,采用基于QR分解的递推最小二乘估计方法实现对采样数据的序贯处理,并求解出待辨识参数。仿真结果表明,提出的激励轨迹设计方法可以显著提高空间机器人参数辨识的收敛速度和准确性。

The problem of dynamic parameter identification of a space robot system after capturing a target is investigated. An excitation trajectory design method is proposed to maintain persistent excitation for the dynamic parameter identification of a free-floating space robot. Firstly, the dynamic parameter identification model of the free-floating space robot is derived based on the momentum conservation principle. Next, finite Fourier series is used to parameterize the joint trajectories and the coefficients of the finite Fourier series are optimized for minimal condition number of the regression matrix as a constrained nonlinear optimization problem. The QR decomposition based recursive least-square algorithm is used to estimate the unknown dynamic parameters. The numerical simulation results show that the proposed excitation trajectory design method is able to improve the convergence rate and accuracy of the parameter identification for free-floating space robots.