Design and verification of a litchi combing and cutting end-effector based on visual-tactile fusion

Abstract

According to the characteristics of litchi tree branching and litchi fruit growing in clusters, this paper designs a litchi combing and cutting end-effector based on visual-tactile fusion with reference to human hair-cutting behavior. The litchi information in the image is perceived visually using the zero-sample prediction large model and the Hough circle detection principle, the tactile timing of the litchi fruit stem is captured by the tactile sensors between the clamping fingers, and the fruit stem clamping status is monitored through the tactile timing update mechanism, and finally the fusion at the decision-making level is realized by integrating visual and tactile in information, realizing the perception of the litchi fruit stem for the litchi picking robot. The picking trial results show that the static tactile fruit stem perception rate is only 16.7%, while the use of visual-tactile fusion increases the rate to 86.7%. The average of a single picking achieves 1.08 bunches. The end actuator and clamping and cutting picking method in this paper provide technical basis for efficient and low-loss litchi picking.