Abstract
This paper proposes the approximation method of evaluation functions of perceived load burden based on the joint moment and investigates the effect of individual difference. Three approximation models (i.e. linear model, exponential model, and logistic function model) are applied to predict the perceived discomfort, and response surfaces of perceived comfort for each participant are obtained. It is found that the average errors of exponential model and logistic function model are significantly lower than those of linear model. In addition, most of the average errors of logistic function are lower than those of exponential model; therefore, the logistic function is selected as the approximation model of perceived discomfort. The amount and location of maximum gradient are defined as the features of the response surfaces, and the correlation coefficients between the features and the height, weight, and gripping force of participants. However, it is found that the number of significant correlation coefficient is very few. Thus, it is perceived that the height, weight, and gripping force cannot reflect the individual variability. The methodology of L-R fuzzy number is utilized so as to consider the individual variability of perceived discomfort, and the response surfaces with its distribution are obtained.
