基于机器学习的SCARA机器人故障诊断方法  

Fault Diagnosis Method for SCARA Robots Based on Machine Learning

作　　者：黄跃珍 明志茂 赵可沦 Huang Yuezhen;Ming Zhimao;Zhao Kelun

机构地区：[1]广州数字科技集团有限公司 [2]广州数据集团有限公司 [3]广电计量检测集团股份有限公司

出　　处：《工程机械》2024年第8期140-145,I0017,共7页Construction Machinery and Equipment

摘　　要：为了提升SCARA机器人故障诊断效果,提出了基于机器学习的SCARA机器人故障诊断方法,利用萤火虫算法(Firefly Algorithm,FA)对支持向量机(Support Vector Machine,SVM)的核参量和惩罚参量进行优化,得到了FA-SVM机器学习方法。利用FA-SVM机器学习法对SCARA机器人进行故障诊断,以机器人振动信号输入量,故障类型为输出量,建立基于FA-SVM的SCARA机器人故障诊断模型。仿真试验结果表明,FA-SVM模型对SCARA机器人故障诊断结果的正确率为100%,验证了所提方法的有效性。To improve the fault diagnosis effect for SCARA robots,a fault diagnosis method for SCARA robots based on machine learning is proposed,and the firefly algorithm(FA)is applied to optimize the kernel parameter and penalty parameter of support vector machine(SVM)to obtain the FASVM machine learning method.The FASVM machine learning method is used to diagnose the faults of SCARA robots,and with the robot vibration signal as the input and the fault type as the output,a fault diagnosis model for SCARA robots based on FA-SVM is established.The simulation test results show that the correctness of the FA-SVM model for SCARA robot fault diagnosis is 100%,which verifies the effectiveness of the proposed method.

关 键 词：SCARA机器人 故障诊断 机器学习 支持向量机 萤火虫算法 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置] TP181[自动化与计算机技术—控制科学与工程]