Analysis of a “Clamp-Pull-Fold” composite end effector for a chili pepper harvesting robot

The fresh edible chili fruits are densely packed, with thin fruit stalks and limited harvesting space. The traditional single stretching or bending method is unable to balance the separation of fruit stalks, the protection of fruits, and the avoidance of collisions between adjacent fruits. To address this issue, this study designed a combined harvesting method of clamping, stretching, and bending, as well as a flexible-rigid coupling end effector. Through the analysis of the force and fracture process of the fruit-stalk-branch system using a flexible contact multi-body dynamics model, and by conducting 16 sets of numerical experiments using a L16 orthogonal design with 5 factors and 4 levels, each combination was not subjected to repeated calculations. Laboratory verification selected fresh edible chili at the same maturity stage, with intact appearance and no obvious defects, and divided them into two categories based on fruit size; the original records did not separately register the variety name, sample total quantity, and the number of repetitions for each category. The determination of success was when the fruit stalks reached the fracture criterion and the fruits were completely separated; failure was determined when the fruits were not fully separated, there was sliding during clamping, or there were visible surface damages. The results showed that under the material parameters and boundary conditions of this study, 7-12 N could be used as a reference range for clamping force, and the combined action of clamping, stretching, and bending could form stress concentration near the fruit stalk separation layer; the optimal parameters obtained within the limited experimental range were a clamping force of 7 N, a stretching speed of 40 mm/s, a stretching displacement of 8 mm, a bending angle of 35°, and a bending angular speed of 90°/s. The average harvesting time of the two types of chili by the laboratory prototype was approximately 3.5 s and 4.0 s, respectively. Due to the limited sample size and repetition times recorded in the physical experiments, it is impossible to calculate a reliable confidence interval. The obtained results only reflect the performance of the device under fixed plant and stable light conditions in the laboratory and still need to be verified through expanding the sample size and conducting multi-variety field repeated experiments to validate its statistical stability and applicability. Robotic arm type Harvesting methods Force shortcoming