Key points are not available for this paper at this time.
Abstract Artificial tactile perception systems that emulate the functions of slow adaptive (SA) and fast adaptive (FA) cutaneous mechanoreceptors are essential for developing advanced prosthetics and humanoid robots. However, constructing a high‐performance sensory system within a single device capable of simultaneously perceiving both static and dynamic forces for surface‐texture recognition remains a critical challenge; this contrasts with common strategies integrating individual SA‐ and FA‐mimicking sensors in multi‐layered, multi‐circuit configurations. Herein, a textile pressure/tactile (PT) sensor is reported based solely on piezoresistive principle alongside high sensitivity and rapid response to both high‐frequency vibrations and static forces. These characteristics are attributed to the sensor's 3D multiscale architecture and the corresponding hierarchical structural deformation of its honeycomb‐like sensing fabric. As a proof‐of‐concept application relevant to humanoid robotics and prosthetics, an automated surface‐texture‐recognition system is constructed by integrating the PT sensor with machine‐learning algorithms, a prosthetic device, an industrial robot arm, and a graphical user interface. This artificial sensory system demonstrates the ability to learn distinct object features, differentiate fine surface textures, and subsequently classify unknown textiles with high recognition accuracy (>98.9%) across a wide range of scanning speeds (50–300 mm s −1 ). These results show promise for the future development of interactive artificial intelligence.
Zheng et al. (Tue,) studied this question.