Guoyang Zhou (PI), Denny Yu
Purdue University,
School of Industrial Engineering
Overexertion
in lifting tasks is one of the leading causes of occupational injuries. Lifting
loads are the key information practitioners require to evaluate the risks of
lifting tasks. However, this is challenging because weight varies across
different objects and is unknown in many circumstances. Existing methods of
predicting lifting loads focused on analyzing body kinematics or muscles’
electrical activities, which are indirect indicators of weight and require
intrusive wearable sensors. This study proposed utilizing tactile gloves
embedded with multiple pressure sensors as a new modality to predict lifting
loads. Hand pressure data measured by tactile gloves during each lift was
formulated as a two-dimensional matrix containing spatial and temporal
information. Different types of deep neural networks were adopted, and a
transferred ResNet 18 regression model achieved the best performance.
Specifically, it achieved a predicted R-squared of 0.821 and a mean absolute
error of 1.579 kg. In addition, to understand the model’s decision-making logic
and the hand force pattern during lifting, the Shapley Additive Explanations
(SHAP) technique was utilized to determine the importance of each sensor at
each frame. The results demonstrated that the right hand was more important
than the left hand for the model to predict lifting loads. Similarly, fingers
were more important than palms, and the middle phase of a lifting task was more
important than its beginning and ending phases. Overall, this study
demonstrated the feasibility of using tactile gloves to predict lifting loads
and provided new scientific insights on hand force exertion during lifting.