Factors affecting human hand grasp type in tomato fruit-picking: A statistical investigation for ergonomic development of harvesting robot

Abstract

The aim of this study is to investigate the factors that affect the human hand grasp type in tomato fruit picking. The correlation between the parameters of the growing environment, fruit size, human body characteristics and choice of grasp type was investigated using a multinomial logistic regression analysis. In tomato fruit-picking, Power palm-Thumb abduction was the most frequently used grasp type, followed by Power palm-Thumb adduction, Power pad-Thumb abduction and Pinch. The plant posture and position of fruits in a tomato plant had no significant effect on the choice of the grasp type during tomato picking (p > 0.05). On the other hand, the major diameter of the fruit, right-hand length, forearm length, middle-finger length, and stature (participant height) had significant contributions on the choice of the grasp type (p < 0.05). With other independent variables unchanged, the probability that Power palm-Thumb abduction was chosen for picking was higher than those of Power pad-Thumb abduction and Pinch when the major diameter of fruits were 1 mm longer; when both the stature and forearm length of participants were 1 cm longer, the corresponding probability that Power palm-Thumb abduction was chosen for picking was higher than those of Power pad-Thumb abduction and Power palm-Thumb adduction, respectively; when both the hand length and middle-finger length of participants were 1 mm longer, the corresponding probability that Power palm-Thumb adduction and Pinch, respectively, were chosen for picking was higher than that of Power palm-Thumb abduction. This study provided an in-depth scientific guidance for dimensional synthesis of fruit-harvesting robots as well as in making grasp planning algorithms for intelligent picking of multi-finger end-effectors from the viewpoint of ergonomics.