Machine Learning Methods to Improve the Accuracy of Industrial Robots
参考中译:提高工业机器人精度的机器学习方法


          

刊名：SAE International Journal of Advances and Current Practices in Mobility
作者：Lauren McGarry（Queen's University Belfast）
Joseph Butterfield（Queen's University Belfast）
Adrian Murphy（Queen's University Belfast）
Colm Higgins（Queen's University Belfast）
刊号：870B0001-99/I
ISSN：2641-9637
出版年：2023
年卷期：2023, vol.5, no.5
页码：1900-1918
总页数：19
分类号：U
语种：eng
文摘：There has been an ongoing need to increase the application of industrial robots to complete high-accuracy aerospace manufacturing and assembly tasks. However, the success of this is dependent on the ability of robotic systems to meet the tolerance requirements of the sector. Machine learning (ML) robot error compensation models have the potential to address this challenge. Artificial neural networks (ANNs) have been successful in increasing the accuracy of industrial robots. However, they have not always brought robotic accuracy within typical aerospace tolerances. Methods that have not yet been investigated to further optimize the use ANNs used in ML robot error compensation methods are presented in this paper. The focus of ML compensation methods has dominantly surrounded ANNs; there have been little to no investigations into other types of ML algorithms for their suitability as robot error compensation models. The success of ANNs to date proves the capability of ML algorithms for this task, and therefore other ML algorithms should be investigated to determine their capability to potentially improve industrial robot accuracy. This paper takes a novel approach by investigating the Support Vector Regression (SVR) ML algorithm to compensate for robot error. The ML models in this research were trained using measurement data captured using a laser tracker and collaborative robot. The ANN model reduced the mean error by 46.4%, 94.8%, and 95.8%, in the x, y, and z-axis, respectively. The SVR model reduced the mean error by 42.4%, 95.9%, and 98.4%, in the x, y, and z-axis, respectively, demonstrating its ability to be implemented as a robotic error compensation model. The success of both the ANN and SVR algorithms enforces the need for further research into other ML algorithms as robot error compensation models, and there is also still potential to further optimize the algorithms used.
参考中译：一直需要增加工业机器人的应用，以完成高精度的航空航天制造和组装任务。然而，这项工作的成功取决于机器人系统满足该部门公差要求的能力。机器学习(ML)机器人误差补偿模型有可能解决这一挑战。人工神经网络(ANN)已经成功地提高了工业机器人的精度。然而，它们并不总是将机器人的精度带到典型的航空航天公差范围内。提出了在最大似然机器人误差补偿方法中进一步优化神经网络使用的方法。最大似然补偿方法的焦点主要围绕神经网络；对于其他类型的最大似然算法作为机器人误差补偿模型的适用性，很少或根本没有研究。到目前为止，人工神经网络的成功证明了ML算法对这一任务的能力，因此应该研究其他ML算法，以确定它们的能力，以潜在地提高工业机器人的精度。通过研究支持向量回归(SVR)最大似然算法来补偿机器人的误差，提出了一种新的方法。本研究中的ML模型是使用激光跟踪器和协作机器人捕获的测量数据进行训练的。神经网络模型在x、y和z轴上的平均误差分别降低了46.4%、94.8%和95.8%。SVR模型在x、y和z轴上的平均误差分别降低了42.4%、95.9%和98.4%，证明了其作为机器人误差补偿模型的能力。人工神经网络和支持向量机算法的成功加强了对其他最大似然算法作为机器人误差补偿模型的进一步研究的需要，而且还存在进一步优化所用算法的潜力。



国家科技图书文献中心  全球文献资源网  京ICP备05055788号-26  机械工业信息研究院 2018-2024