A Printing Pressure expectation system based on fuzzy inference has been installed in a liquid crystal display panel (LCD) production plant to solve problems in determining the printer printing pressure in real-world LCD production, in which the recognition of printing pressure conditions and control are very important and difficult, influencing product quality. This is usually done conducted by skilled engineers, whose performance depends on tacit knowledge. We propose using fuzzy inference to solve this problem. Printing area images are observed with cameras and abstract features extracted using image processing. System output is the status of printing pressure, divided into excessive pressure (EP), good pressure (GP), and low pressure (LP). Based on abstract features, the state is calculated using fuzzy membership functions. Shapes of membership functions are determined based on sampled glass obtained in actual LCD production line. Experiments are conducted using 2000 samples of glass printed using actual printers, or which results are compared to those of skilled engineers. We found that the proposed system yields quality higher than that of skilled engineers. We installed our system on an actual production line, where it is expected to increase product quality and production speed while and cutting production costs.

1. [1] J. Jianping, Y. Takama, and T. Yamaguchi, “A Proposal of Printing Pressure Control System based on fuzzy Inference Method,” 6^th Int. Symposium on Advanced Intelligent System (ISIS2005), pp. 355-360, Yosu, Korea.
2. [2] K. Hirota, N. Iwamatsu, and Y. Takama, “Proposal of an Internal-State Inference System based on Multimodal Sensory Fusion Method,” IEEE Transactions on Instrumentation and Measurement, Vol.51, No.2, pp. 347-352, 2002/4.
3. [3] K. Hirota, N. Iwamatsu, and Y. Takama, “Inspection System based on Fuzzy Multi-modal Sensory Integration,” 2nd Int. Symposium on Advanced Intelligent System Conference (ISIS2001), pp. 109-113, 2001/8.
4. [4] K. Hirota, N. Iwamatsu, and Y. Takama, “Toward the Realization of Multimedia Instrumentation based on Soft Computing Technology,” Japan-Tunisia Workshop of Informatics (JTWIN2001), pp. 90-95, 2001/10.
5. [5] K. Hirota, Y. Takama, and N. Iwamatsu, “Multimedia Nondestructive Inspection based on Computational Intelligence Technology (Opening Lecture),” 2nd Int. Conf. on Intelligent Technologies (INTech2001), (Bangkok Thailand), pp. 14-20, 2001/11.
6. [6] K. Hirota, “Fuzziness Intelligence Robot Control,” Chapter 2, Mc-Graw, Hill Book (Tokyo), 1985.
7. [7] T. Terano, K. Asayi, and M. Kanno, “Fuzzy Systems Theory and Its Applications,” Ohm Publish,
   ISBN4-274-02124-6,
   1987.
8. [8] W. E. Snyder and H. Qi, “Machine Vision,” Cambridge University Press,
   ISBN 0-521-83046-X,
   2004.
9. [9] R. O. Duda, P. E. Hart, and D. G. Stork, “Pattern Classification,” New Technology Communications Co., Ltd.,
   ISBN 0-471-05669-3,
   2001.