Accuracy of Estimating Hand Location During Lifting Using Five Wearable Motion Sensors

Abstract

The purpose of this study was to assess two computation models for estimating the hand locations during lifting tasks using data from five inertial measurement units (IMUs) attached to five body segments. The first model computed the hand location with the IMU gyroscope data and the pre-defined ratios of body segment lengths. The second model used the same gyroscope information and all measured lengths of the body segments. The outcome measure of these models was the estimated hand location in 12 lifting zones defined by the ACGIH Threshold Limit Values (TLVs) for lifting. Motion data was collected with the wearable system and a laboratory-grade motion capture system on ten subjects that performed 12 two- handed lifting tasks representing the lifting zones. By including body segment measurements, the average accuracy of the model improved from 4 to 34%, suggesting that body segment information plays an important role in estimating the lifting zones.