Optimization of a bionic end-effector for automated cherry tomato harvesting using Response Surface Methodology and Box-Behnken design

Harvesting cherry tomatoes is a labor-intensive and time-consuming endeavor. The implementation of robots for this task represents an effective solution. The end-effector is a vital component of the harvesting robot, essential for automated picking of cherry tomatoes. This study introduces an end-effector with three grippers, modeled on the structure and mechanics of the phalangeal chain found in vultures' legs, to enhance the efficiency of harvesting robots. To optimize the bionic end-effector's structural parameters for cherry tomato picking, Response Surface Methodology (RSM) with a Box Behnken Design (BBD) and a desirability-based multi-response optimization strategy was employed. Experiments varied four process parameters at three levels: Width of the Distal Phalanges and Middle Phalanges (WDMP), Thickness of the Distal Phalanges and Middle Phalanges (TDMP), Angle at the End of the Distal Phalanx (AEDP), and Distance between the Heel Phalanges and the Central Axis of the Fixed Seat (DPAS). The maximum inscribed circle diameters of the middle phalanges (DCMP) and distal phalanges (DCDP) were selected as surrogate geometric responses because they represent, respectively, the compressive clearance related to fruit damage risk and the distal enclosure space related to fruit retention stability. The optimal configuration was identified as WDMP 16 mm, TDMP 5 mm, AEDP 90°, and DPAS 7.5 mm. Furthermore, experimental validation of the optimized bionic end-effector demonstrated a harvesting success rate of at least 88% within the tested equatorial-diameter range. This research provides valuable insights and a theoretical basis for future designs of end-effectors for cherry tomato harvesting.