ABSTRACT: 

Flexible tactile sensors equipped with multimodal decoupling capabilities are essential for the advancement of intelligent robotics. However, these sensors face challenges related to signal interference and limitations in single-variable detection. Drawing inspiration from the multilayered architecture and perception ability of biological skin, this study presents the development of an asymmetric gradient WPU-CNT@Bi2Te3 film via a gravity-induced self-sedimentation strategy, which integrates piezoresistive and thermoelectric effects, facilitating a dual-mode decoupled sensing capacity. The gradient distribution of CNT@Bi2Te3 within the polyurethane matrix significantly enhances the bending sensitivity and thermoelectric output by establishing differential pathways for strain fields and reducing the thermal conductivity. When coupled with flexible heaters, the as-derived flexible tactile sensor can achieve an ultrahigh recognition accuracy of 100% in classifying complex objects possessing various shapes and materials simultaneously by using a multimodal fusion CNN algorithm. Moreover, this sensor demonstrates exceptional stability and conformal adaptability, thereby enabling a robotic hand with precise grasping ability.

URL: https://pubs.acs.org/doi/10.1021/acs.nanolett.5c03246