OPTIMAL CALIBRATION AND OPTIMAL DESIGN OF HIGH PRECISION PARALLEL MECHANISMS

为了实现高精度的并联机构，需要进行精确的运动学标定，并考虑标定的影响进行优化设计。对于精确的运动学标定，我建议使用通过对周期性测量路径(如圆形路径)进行数据变换而获得的低阶傅立叶级数。提出了两个无量纲指标来评价测量路径和校准参数的独立程度。还提出了一项关于测量误差的指标来评价校准的准确性。文中还提出了一种确定最优测量路径和使用这些指标的指定路径数的最优集合的算法。利用自行研制的六自由度空间并联机构，在实验工作台上进行了仿真和实验，验证了所提出的方法、指标和算法的有效性。利用不同运动常数并联机构的仿真结果，讨论了考虑运动学标定影响的优化设计问题。通过仿真研究了输出位移误差的测量策略对运动学标定结果的影响。

In order to realize a high precision parallel mechanism, it is necessary to carry out precise kinematic calibration and to carry out optimal design with consideration of effect of calibration. For precise kinematic calibration, I proposed using a low order Fourier series obtained by transforming the data for periodic measurement paths such as circular paths. Two non-dimensional indices were proposed for evaluating the degree of independence regarding the measurement paths and calibration parameters. An index was also proposed to evaluate the accuracy of the calibration with regard to measurement error. An algorithm for determining optimal measurement paths and an optimal set of a specified number of paths using these indices was also proposed. The effectiveness of the proposed method, indices, and algorithm was investigated through simulations and experiments with an experimental worktable using a spatial parallel mechanism with six degrees of freedom that was developed in my laboratory. Using simulation results of parallel mechanisms with different kinematic constants, optimal design with consideration of the effect of kinematic calibration was discussed. The effect of measurement strategies of output displacement error on the result of kinematic calibration was investigated through simulations.