TY - JOUR
T1 - An automated dynamic-balancing-inspection scheme for wheel machining
AU - Tieng, Hao
AU - Li, Yu Yong
AU - Tseng, Kuang Ping
AU - Yang, Haw Ching
AU - Cheng, Fan Tien
N1 - Funding Information:
Manuscript received September 10, 2019; accepted January 13, 2020. Date of publication February 3, 2020; date of current version February 17, 2020. This letter was recommended for publication by Associate Editor Prof. C.-B. Yan and Editor Prof. J. Yi upon evaluation of the reviewers’ comments. This work was financially supported by the “Intelligent Manufacturing Research Center” (iMRC) from The Featured Areas Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education (MOE) in Taiwan, ROC. This work was also supported by the Ministry of Science and Technology of Taiwan, ROC under Grants MOST 108-2221-E-006-210-MY3 and MOST 108-2218-E-006-029. (Corresponding author: Fan-Tien Cheng.) H. Tieng, Y.-Y. Li, K.-P. Tseng, and F.-T. Cheng are with the Institute of Manufacturing Information and Systems, National Cheng Kung University, Tainan, Taiwan 71147, ROC (e-mail: [email protected]; yuyung1104@imrc. ncku.edu.tw; [email protected]; [email protected]).
Publisher Copyright:
© 2016 IEEE.
PY - 2020/4
Y1 - 2020/4
N2 - Wheel balance plays an important role in vehicle safety. The existing inspection method for wheel balance mainly relies on the off-machine measurement technique, which is time-and manpower-consuming as the worldwide requirement of the automated production system gradually increases. However, the multi-unbalance causes are difficult to identify due to complex machine structures; and the low signal-noise-ratio between wheel and machine vibration makes traditional handcrafted features difficult to detect wheel unbalance. To overcome these two challenges, this paper proposes a Dynamic-Balancing-Inspection (DBI) scheme which integrates steps of data collection, data preprocessing, and ensemble average of Convolution Neural Network (CNN) based models with well-Tailored filters and activation functions, to automatically uncover critical information from frequency data and provide the on-machine and real-Time total inspection for the wheel balance. The application of the wheel balance from a practical CNC-machine is adopted to illustrate the performance of the DBI approach.
AB - Wheel balance plays an important role in vehicle safety. The existing inspection method for wheel balance mainly relies on the off-machine measurement technique, which is time-and manpower-consuming as the worldwide requirement of the automated production system gradually increases. However, the multi-unbalance causes are difficult to identify due to complex machine structures; and the low signal-noise-ratio between wheel and machine vibration makes traditional handcrafted features difficult to detect wheel unbalance. To overcome these two challenges, this paper proposes a Dynamic-Balancing-Inspection (DBI) scheme which integrates steps of data collection, data preprocessing, and ensemble average of Convolution Neural Network (CNN) based models with well-Tailored filters and activation functions, to automatically uncover critical information from frequency data and provide the on-machine and real-Time total inspection for the wheel balance. The application of the wheel balance from a practical CNC-machine is adopted to illustrate the performance of the DBI approach.
UR - http://www.scopus.com/inward/record.url?scp=85080953942&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85080953942&partnerID=8YFLogxK
U2 - 10.1109/LRA.2020.2970953
DO - 10.1109/LRA.2020.2970953
M3 - Article
AN - SCOPUS:85080953942
SN - 2377-3766
VL - 5
SP - 2224
EP - 2231
JO - IEEE Robotics and Automation Letters
JF - IEEE Robotics and Automation Letters
IS - 2
M1 - 8978473
ER -