Abstract
The aim of this study was to formulate a relationship between the total perceived discomfort of the upper limbs and the perceived discomforts of each degree of freedom (DOF). The biomechanical analysis was performed based on classifying the posture and load via rapid upper limb assessment (RULA) that is an observational method used to assess the total physical workload. The joint moments and the perceived discomforts for each DOF were calculated. The summary scores of the RULA and the perceived discomforts of each DOF were the objective and explanatory variables, respectively. Four approximation models (average, maximum, hybrid of average and maximum, and radial basis function (RBF) network) were compared in terms of accuracy in predicting the total perceived discomfort. The results indicated that the dominating factor for the total discomfort transitions from the average to maximum discomfort depending on the maximum discomfort of each DOF. In conclusion, among the four models, the RBF network provided the highest accuracy, followed by the hybrid model.
