The robot calibration exists many excellent performances by using the POE formula, but it demands extra works to deal with the problems of the joint twist constrains. Thus, the Adjoint error model is proposed to solve these problems, which can be developed further. In this paper, we have modified the Adjoint error model and improved the calibration method. On the one hand, we conduct the formula entirely for the serial robot based on the modified Adjoint error model, so as to give an explicit and unified calibration model. On the other, the method of updating kinematic parameters is also simplified, comparing the previous calibration methods with the Adjoint error model, which can improve the convergence rate and optimize our calibration process. For verifying this method, the simulations on the typical robots of PUMA 560 and SCARA are presented, and the results show that this method has evident superiorities. GRAPHICAL ABSTRACT

中文翻译：

---

一种改进的基于POE的伴随误差模型的机器人标定方法

使用POE公式进行机器人标定有很多优异的性能，但需要额外的工作来处理关节扭曲约束的问题。因此，提出伴随误差模型来解决这些问题，并可以进一步发展。在本文中，我们修改了伴随误差模型并改进了校准方法。一方面，我们在改进的伴随误差模型的基础上，对串行机器人进行了完整的公式计算，从而给出了一个明确统一的标定模型。另一方面，运动学参数的更新方法也得到了简化，将以往的标定方法与伴随误差模型进行比较，可以提高收敛速度，优化我们的标定过程。为了验证这种方法，给出了 PUMA 560 和 SCARA 典型机器人的模拟，结果表明，该方法具有明显的优越性。图形概要