Purpose

Every geometric model corresponding to a unique feature whose errors of parameters uncorrelated, so the linearization technique can be successfully applied. The solution of a linear least square problem can be applied straightforwardly. This method has advantages especially in calibrate the redundant robot because it’s relatively small. The parameters of kinematics are unique and determined by this algorithm.

Design/methodology/approach

In this paper, a geometric identification method has been studied to estimate the parameters in the Denavit–Hartenberg (DH) model of the robot. Through studying the robot’s geometric features, specific trajectories are designed for calibrating the DH parameters. On the basis of these geometric features, several fitting methods have been deduced so that the important geometric parameters of robots, such as the actual rotation centers and rotate axes, can be found.

Findings

By measuring the corresponding motion trajectory at the end-effector, the trajectory feature can be identified by using curve fitting methods, and the trajectory feature will reflect back to the actual value of the DH parameters.

Originality/value

This method is especially suitable for rigid serial-link robots especially for redundant robots because of its specific calibration trajectory and geometric features. Besides, this method uses geometric features to calibrate the robot which is relatively small especially for the redundant robot comparing to the numerical algorithm.

中文翻译：

---

基于几何模型的刚性串行链接机器人DH参数识别新方法

目的

每个几何模型都对应于其参数误差不相关的唯一特征，因此可以成功地应用线性化技术。线性最小二乘问题的解可以直接应用。该方法相对较小，因此在校准冗余机器人方面特别具有优势。运动学的参数是唯一的，并由该算法确定。

设计/方法/方法

在本文中，研究了一种几何识别方法来估计机器人的Denavit-Hartenberg（DH）模型中的参数。通过研究机器人的几何特征，设计了特定的轨迹来校准DH参数。基于这些几何特征，推导了几种拟合方法，以便可以找到机器人的重要几何参数，例如实际的旋转中心和旋转轴。

发现

通过在末端执行器上测量相应的运动轨迹，可以使用曲线拟合方法识别轨迹特征，并且该轨迹特征将反映回DH参数的实际值。

创意/价值

由于其特定的校准轨迹和几何特征，此方法特别适合于刚性串行链接机器人，尤其是冗余机器人。此外，该方法利用几何特征来校准机器人，该机器人相对于数值算法而言相对较小，特别是对于冗余机器人而言。