Bioinspired Integrated Multidimensional Sensor for Adaptive Grasping by Robotic Hands and Physical Movement Guidance

Abstract

Listen

Translate article

AbstractThe construction of piezoresistive sensors capable of distinguishing multiple mechanical stimuli is important for sensing higher level applications. However, the mutual interference between sensing signals is a technical bottleneck for sensors to recognize multiple mechanical stimuli. Inspired by the structure of human skin and muscle, a multidimensional sensor is designed by integrating three sub‐sensors. Each sub‐sensor is only sensitive to stimulus components in one of the three orthogonal axes, whereas it remains insensitive to others because of its anisotropic sensing properties. Specifically, an omnidirectional gradient wrinkled polyurethane film with MXene‐embedded ZnO nanowire arrays serves as a strain‐insensitive pressure sub‐sensor, while two aligned segmental polyimide/polyurethane films through orthogonal stacking act as two pressure‐insensitive anisotropic strain sub‐sensors. A unique characteristic of multidimensional sensor is its ability to distinguish and measure the type, magnitude, and direction of strain, pressure, and shear. Moreover, multidimensional sensor exhibits maximal gauge factor of 863.7 for strain and highest sensitivity of 187.71 kPa−1 for pressure. Importantly, multidimensional sensor exhibits an unprecedented ability to quantitatively evaluate shear using a library of electrical responses. The adaptive grasping of robotic hands and free‐throw training of players have been demonstrated to initiate the development of robotic object manipulation and physical movement guidance.