Bio-Inspired Soft Hand with Cognitive Touch for Fruit Sorting

The hand is a highly integrated biomechanical system that combines rigid bones, flexible joints, and compliant soft tissues to achieve precise and adaptive motion. Its layered structure, enhanced by tactile skin and mechanoreceptors, enables delicate manipulation tasks. Inspired by this synergy, we present a bio-inspired soft robotic hand that emulates these principles through muscle-like actuation and embedded sensing. The design incorporates PUY-based strain sensors within soft actuators, enabling real-time detection of bending angles and contact forces. By analyzing deformation responses to varying textures and resistances, the system infers key fruit characteristics such as size and firmness—critical indicators of ripeness and type. Experiments demonstrate reliable classification of different fruits and ripeness levels, highlighting the potential for adaptive, compliant manipulation in agricultural and food handling applications. Future work will expand the grasping dataset, enhance classification accuracy for delicate or irregular objects, and explore biomedical applications for automated quality assessment of fragile items. © 2025 IEEE.