Spatially Selective and Stepless Human-Intention Control Framework for Variable Stiffness and Grasp Force of Soft Grippers

Abstract

The use of soft robotic end-effectors has been prevalent in sensitive agricultural harvesting because these devices have high compliance and are safe to interact with humans. Current variable-stiffness grippers have the disadvantages of full-finger uniform stiffness parameters and discrete-grade parameter control, which can cause mechanical shock and soft fruit damage easily. The existing brain-computer interface (BCI) control schemes can only give out binary control commands and are incapable of affording continuous adjustment due to human grasping perception. To overcome these challenges, the current paper has suggested a human-intention control system using regional regulation and output that is stepless with multi-segment soft gripper. A two-mode BCI that incorporates motor imagery (MI)-based offline calibration and error-related potential (ErrP)-based online supervision is assembled to reach the segmented stiffness tuning and full-range continuous grasping force control. BCI anomaly monitoring and hierarchical dual closed-loop control with tactile feedback enhances environmental adaptability in unstructured farm environments. The research offers a new human-involved smart handling approach to variable stiffness soft manipulators and improves nondestructive harvesting abilities of fine agricultural farming.