반복적인 RSM과 RLS를 이용한 LCD 모듈 실장 설비의 모델 추정 및 정도 품질 향상

Abstract

Recently mounting technology with high specifications is required for of high-precision LCD TV. In the LCD module manufacturing process, the OLB process includes vision system recognizing the distance between electronic device and LCD pattern, linear motion system adjusting alignment with a backlash, bonding system with high temperature and pressure, and finally, inspection system. The vision system with lens distortion and resolution is modeled as a 3-order polynomial. The motion system model consists of a conditional backlash. The bonding system model contains a few rational expressions with lot-based CTE. Also each model has standard normal distribution noise. Unknown parameters of the models are estimated with input and output data using recursive RSM and RLS. R-square and RMAE are used to measure the accuracy of the estimated parameters. The estimated offset and setting parameters from the reliable model improve mounting accuracy. The trend of estimated parameters helps monitoring and diagnosis of equipment status. This proposed algorithm has been tested with simulation.