An anthropometric study for the anthropomorphic design of tomato-harvesting robots

Abstract

To explore how humans can dexterously pick fruits of various sizes from a tomato plant from different positions and postures, some anthropometric dimensions related to the hand-harvesting of the fruits were determined within 261 volunteers, and their correlations were investigated using a bivariate correlation analysis method. The heights of the body (stature), shoulder tip, waist and knee of participants ranged from 111.2 to 193.6, 87.5 to 163.0, 62.1 to 117.3, and 31.0 to 56.0 cm, respectively; the ratios of the knee height, waist height, and shoulder tip height to the stature were 0.25–0.33, 0.52–0.65, and 0.76–0.87, respectively; the shoulder tip breadth, upper limb length and foot length of participants ranged from 26.0 to 48.5, 45.4 to 85.6, and 19.5 to 28.0 cm, respectively; and the length ratios of the hand, forearm, and upper arm to the upper limb were 0.23–0.31, 0.29–0.38, and 0.35–0.45, respectively. There were high linear correlations among the stature, waist height and knee height, between the forearm and hand lengths, and among the stature, hand length and arm length and moderate linear correlations between most hand dimensions. These human body dimensions determine the upper limb workspace, body balance, working comfort and lower limb movement frequency during picking, and this information will be useful for the bionic dimensional synthesis design of a harvesting robotic arm, walking mechanism and body. The results of multiple comparisons in ANOVA showed that there were significant differences between some of the finger dimensions and between some of the hand-scaling dimensions, which is one of the main features enabling the human hand to pick various sizes of fruits. This study can provide guidance and inspiration for the anthropomorphic design of robotic tomato-harvesting end-effectors.