A FUZZY-LOGIC ANALYSIS METHOD FOR EVALUATING HUMAN SENSITIVITIES

As a basic research tool for modeling human sensitivity, we have devel oped a program for analyzing sensitivity evaluation data using a fuzzy regression method. This takes into account the ambiguities and nonlin earity of human sensations and thus provides a method for quantifying their correlations with physical characteristics. This was applied to a study of thermal sensation in a vehicle interior. Optimum membership functions were obtained that show the nonlinear relationship between each physical variable (e.g., air flow velocity, temperature and sun l oad) and thermal sensation. When a fuzzy regression analysis was perfo rmed using the membership functions, it was found that 85% of the eval uation scores that human subjects gave to various thermal sensations w ere included within the range of the estimated values.