Nonlinear Predictive Functional Control Based on Support Vector Machine

Yang, Xiyun; Lv, Yuegang; Xu, Dehong; Zhang, Xinfang

doi:10.1109/icmlc.2006.258517

Cited by 1 publication

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…OF SUPPORT VECTOR MACHINE Support vector machine is a machine learning method based on VC dimension theory in statistical learning theory and structural risk minimized principle, which can find the best compromise between complexities of model and learning ability through limited sample information, as well as it is a successful accomplishment of SLT [6].…”

Section: IIImentioning

confidence: 99%

Study on a method of dispensing calibration for photoresist pump based on support vector machine

Zhang

Cao

et al. 2015

2015 IEEE International Conference on Cyber Technology in Automation, Control, and Intelligent Systems (CYBER)

View full text Add to dashboard Cite

Photoresist pump is an important part of coater and developer equipment. High precision control for the amount of photoresist is of great significance to improve the quality of coating and reduce the cost of production. Its working principle is dispensing and reloading photoresist by the positive and negative rotation of its internal stepping motor, and the dispensing amount is determined by the pulse count that the pump controller sends. But due to the linearity error of the pump, viscosity of photoresist, damping of the pipeline, pressure of the outlet and factors such as environmental temperature and humidity, there is a deviation between the dispensing actual value and set value, so calibration is needed. For the above question, firstly, in this paper the support vector machine model of the actual dispensing amount is established by dispensing speed and set value as its input vector construction. Secondly, the dispensing revised value is solved by iterative method, and then the calibration coefficients are calculated. Finally, the validity of this method is verified by experimental results, moreover, the dispensing precision is improved and the time which determining calibration coefficients required is reduced by this method.

show abstract

Section: IIImentioning

confidence: 99%