Design optimization and simulation of a flexible manipulator for blueberry picking

To address high damage rates and low adaptability in harvesting delicate lowbush blueberries from dense clusters, this paper designs a rigid-flexible coupled three-fingered picking end-effector driven by a lead screw transmission and linkage mechanism. First, a contact mechanics model was constructed and validated via ANSYS finite element analysis. The maximum contact deformation and surface stress were 0.11 mm and 0.03 MPa, which remained well below the fruit’s respective allowable limits of 1.25 mm and 0.3–0.5 MPa, confirming non-destructive picking feasibility. Second, a kinematic model and an evaluation index system—comprising mean manipulability, trajectory curvature smoothness, workspace radial distance, and trajectory vertical fluctuation—were established to quantitatively assess obstacle avoidance and motion stability in dense environments. Third, structural parameters were globally optimized using a multi-objective optimization model and the NSGA-II genetic algorithm. The optimized manipulator achieved a 53.5% improvement in mean manipulability and a 59.7% increase in trajectory smoothness, effectively minimizing disturbances to adjacent fruits. Lastly, continuous harvesting experiments with 200 trials on a physical prototype demonstrated a 92.0% success rate, a 2.0% damage rate, and a 5.2 s average cycle time. Statistical analysis confirmed high consistency, with a success rate coefficient of variation of 5.89%. These results establish the proposed manipulator as a highly reliable, adaptable, and robust solution for continuous, non-destructive blueberry harvesting.