Parallel robots with R-(SS)² (R, actuated revolute joint; S, spherical joint) branches are widely used in high-speed sorting. This kind of robot has been developed into a variety of robots with different moving platforms, but the error model is usually only applicable to the specific features of the moving platform. Thus, an adaptable error modeling method is in need. To solve this problem, this paper improves the existing error modeling method by considering the error transmission mode of the dual-moving platform as a common straight line. On this basis, the general form of the error model containing all possible geometric errors is established. Based on the error model and contribution concept, sensitivity indices are defined and the sensitivity is analyzed. Then some atlases are obtained to describe each geometric error’s influence on the moving platform’s pose errors. Accordingly, the linear errors and angular errors that have greater impact on the moving platform’s pose accuracy are identified. Compared with the existing error modeling methods, this improved error modeling method takes the adaptability of the error modeling into account, and can be extended to the specific error modeling of the corresponding robot.

带有R-（SS）²（R的并联机器人，致动旋转接头；S，球形接头）分支广泛用于高速分选。这种机器人已经发展成为具有不同移动平台的各种机器人，但是误差模型通常仅适用于移动平台的特定功能。因此，需要一种自适应的错误建模方法。为了解决这个问题，本文以双运动平台的误差传递方式为一条公共直线，对现有的误差建模方法进行了改进。在此基础上，建立了包含所有可能的几何误差的误差模型的一般形式。基于误差模型和贡献概念，定义了灵敏度指标并分析了灵敏度。然后获得一些地图集，以描述每个几何误差对移动平台的姿态误差的影响。因此，识别出对移动平台的姿势精度有较大影响的线性误差和角度误差。与现有的错误建模方法相比，这种改进的错误建模方法考虑了错误建模的适应性，并且可以扩展到相应机器人的特定错误建模。