Compliant picking control of dragon fruit picking robot based on adaptive variable impedance

Abstract

In response to the challenges of picking dragon fruit and the mechanical damage caused to the fruit by picking robots, this study developed a control system for the dragon fruit picking robot. A compliant picking control method for dragon fruit was proposed, which includes a picking strategy and an impedance control algorithm. The picking strategy utilises force sensors to improve the positioning ability of the fruit stalk and picking effect of the manipulator. Based on impedance control theory, a contact force model between the robot and the orchard environment was established, and the adaptive variable impedance control algorithm was improved. The contact force is used as the feedback of the adaptive variable impedance control, and the damping parameters are adaptively adjusted to adjust the contact force between the robot and the orchard environment, thereby reducing potential damage caused by abnormal picking. The picking trial results indicated that the overshoot of the robot under adaptive variable impedance control was 4.53%, with an adjustment time of 3.54 s, reducing by 3.83% and 0.77 s compared to traditional impedance control. The average time to pick a single dragon fruit under adaptive variable impedance control was 36.07 s, with an average success rate of 68.69% and a relatively low damage rate of 20.5%. The designed dragon fruit picking robot is capable of successfully detecting and adapting to occurring anomalies to minimise damage, effectively lowering the damage rate and improving the picking success rate, providing theoretical and technical support for the safe and compliant picking of dragon fruit.