Neuromorphic Tactile Perception Enabled by Triboelectric Artificial Synapse for Material Identification

Abstract With the rapid advancement of intelligent robotics and automated inspection systems, material identification technology faces critical requirements for enhanced precision and adaptive capabilities. However, conventional perception systems struggle to achieve human‐like adaptive learning capabilities through sustained interactions. Owing to their bio‐inspired perceptual characteristics and dynamic optimization mechanisms, neuromorphic devices have garnered significant attention in addressing these technological challenges. Here, this investigates neuromorphic tactile perception enabled by a triboelectric artificial synapse (TAS) for material identification. The TAS utilizes contact‐induced material‐specific triboelectric potentials to drive directional ions migration in ion gels, dynamically modulating semiconductor channel conductance. This modulation accumulates with repeated stimuli. By integrating a convolutional neural network to extract multi‐scale features from conductance response curves, the system exhibits a “contact‐learning‐optimization” paradigm, autonomously enhancing the mapping between material dielectric properties and triboelectric potentials. Experiments show that increasing contact frequency from 2 to 10 within 2s raises recognition accuracy from 86.11% to 97.78%. This work pioneers a TAS‐enabled material identification framework with progressive learning capabilities, establishing a new pathway for developing self‐adaptive intelligent perception systems. The demonstrated paradigm of TAS shows broad application potential across intelligent robotics, adaptive human‐machine interfaces, and industrial automation with environmental variability.