Increasing the sensitivity of posture assessment using rapid upper limb assessment, Bayesian network and fuzzy numbers

This study introduces a novel scoring system based on triangular fuzzy numbers (TFNs) and proposes a new scoring combination approach integrating the original rapid upper limb assessment (RULA) worksheet with a Bayesian Network (BN) model. The resulting method, named fuzzy Bayesian network-based RULA (FBnRULA), was evaluated by comparing it with the RULA method for assessing 14 different postures. The FBnRULA method showed a maximum difference of 1.75 and an average difference of 0.62 compared to RULA. This enhancement addresses the shortcomings of traditional methods that lack sensitivity to variations in input variables, which can hinder the accurate assessment of ergonomic interventions. The developed FBnRULA method demonstrates its effectiveness in detecting minimal changes in ergonomic postures following interventions in industrial settings.