Development of Biomechanical Models for Describing Hand and Finger Placements in Handling Work Objects.

Abstract

A conceptual model was proposed to describe the overall relationship among hand postures, motions, forces, factors, memory, and feedback. Second, logistic regression models were developed based on a study of 10 male and 10 female subjects that showed relative hand load greater than 34% of maximal strength motivated subjects to reach and grasp cylinders using underhand posture (rather than overhand), and relative hand load as low as 24% motivated subjects to hold the objects (for about 8s) using palm grip at shoulder height (vs. hook grip at mid-thigh height). The relative hand load threshold increased to 53% for selecting underhand over overhand posture for placing the objects. Third, a study of relative finger loads for 6 male and 6 female subjects lifting cylinders showed that selection of hand posture is related to the preference of reducing thumb and fingertip forces and joint loads. Subjects demonstrated strong preferences of underhand over overhand grasp, and hook grip over pinch to lift cylinders, while thumb tip and sum of fingertip forces can be reduced up to 60% by selecting the preferred postures. Biomechanical models predicted overhand thumb and finger normal forces similar to data if friction was considered, while predicting about 2 times of measured normal forces if friction force was assumed zero. Fourth, finger force distribution and placement were determined for 6 males and 6 females holding unbalanced plate objects. The thumb and finger center-of-force (CoF) locations were generally aligned with the load moment arm. The distance between thumb and finger CoF locations increased by 39% as load moment increased from 0.98 Nm to 2.35 Nm, and reduced by 17% as hand length increased from 16.2 cm to 21.1 cm when the plate was held horizontally. Previous studies showed that posture selection is related to effort (Rosenbaum et al. 2006). This work shows that effort can be described quantitatively by relative joint loads, and that posture predictions based on biomechanical analysis of relative finger forces and joint loads account for 45%-87% of variance. The unexplained variance may be due to mechanical inter dependencies among finger motions and finger force measurement errors.