QUALITATIVE AND QUANTITATIVE FUZZY INFERENCE METHOD FOR CONSIDERATION OF DISPERTION OF DATA

摘要

Since consumer products continue to move to higher functionality, the need for high-precision technology is increasing. In adjustment process, the problem is that the characteristics of product are different in each work pieces. To handle the dispersion of characteristics of controlled system, we propose a new fuzzy inference method utilizing probability density functions of dispersion of measurement values defined as membership functions. In addition, to control more precisely by consideration of qualitative knowledge of the causes of data dispersion, we propose a method to determine the membership functions using the distributions of grade accumulation. We applied these proposed methods to automatic adjustment device for magnetic head, and the experimental results were satisfactory.Keyword: fuzzy; membership function; probability density function; qualitative knowledge