Vibrissae-inspired wearable neuromorphic artificial fingernail for vibrotactile perception and interfacing

Abstract The rodent vibrissal system exhibits remarkable vibrotactile perception, enabling the discrimination of fine surface features through dynamic spike encoding and efficient synaptic integration. Here, we develop a vibrissae-inspired wearable neuromorphic artificial fingernail that combines piezoelectric sensing, spike encoding, and synaptic processing within a near-sensor computing architecture. A dual-gate synaptic transistor gated by proton-conducting ion-gel exhibits excellent air stability, sensory memory, sensory integration, and perceptual weighting capabilities, allowing spatiotemporal recognition of both fast- and slow-adapting tactile spikes. Notably, this synaptic device features multiple preprocessing functions, enabling direct extraction of task-relevant features at the hardware level and thereby minimizing computational requirements. Using this system, we achieve over 90% classification accuracy across multiple tactile tasks, including object contour recognition, surface texture discrimination, and finger-writing trajectory identification. These results demonstrate the potential of our system as a smart wearable interface for low-power, real-time tactile recognition, thereby expanding the applications of neuromorphic electronics.