Grasping perception and prediction model of kiwifruit firmness based on flexible sensing claw

Fruit firmness is an important quality indicator that affects fruit storage and transportation. Robot automation sorting is becoming a trend in fruit detection and sorting. During the kiwifruit picking and feeding stages, a non-destructive tool was developed that utilizes flexible claws to grasp and sense the firmness of the kiwifruit. This innovative tool aims to enhance the level of intelligence within the kiwifruit industry. In this study, a flexible sensing claw was designed for grasping and perceiving kiwifruit firmness with force sensors and bending sensors. According to the force–deformation law of flexible claw in the process of grasping kiwifruit, a tactile-based firmness index H based on the flexible sensing claw with fruit as six elements Kelvin-Burgers coupling was innovatively proposed as a reference index for firmness prediction. The correlation between the tactile-based firmness index of flexible sensing claw and the reference firmness of kiwifruit was analyzed by creep and puncture tests. Finally, the firmness date and grasping data of 210 kiwifruits were collected experimentally during different storage periods. The tactile-based firmness index of flexible claw can effectively distinguish kiwifruit with different firmness through comparative analysis. And the regression model achieved an approximate prediction for kiwifruit firmness in the external prediction set (RP2 = 0.813, REMEP=0.369 N/mm, RPDP=2.342). The flexible sensing claw system established in this study can characterize the firmness of kiwifruit during picking and placing the fruit.