A Self‐Powered Optical Fiber Tactile Sensor with Crosstalk‐Free Imaging and Human‐Level Sensitivity

ABSTRACT Artificial tactile sensors aim to mimic human touch with high sensitivity, fine spatial resolution, low detection thresholds, and fast response. However, developing a human‐like haptic system that integrates crosstalk‐free large‐scale arrays, ultra‐low power consumption, and efficient signal transmission and interpretation remains a formidable challenge. Inspired by human skin's layered structure and neural pathways, we developed a skin‐like optical fiber tactile sensor. A finger‐like multilayer composite is integrated onto the endsurface of a dense optical fiber array. Using self‐powered mechanoluminescent (ML) materials, this tactile photonic skin visualizes mechanical stimuli in real time, bridging tactile sensing and visual perception. Each optical fiber independently transmits localized luminescent signals, ensuring crosstalk‐free, high‐fidelity spatial encoding. Coupled with an area‐array CMOS imager and vision‐based signal processing algorithms, the system achieves human‐fingertip‐level performance: 7 kPa detection threshold, 0.4 mm spatial resolution, 86 ms response time, and high durability. Demonstrations include optical recognition of alphabetic patterns and screw threads, as well as accurate palpation‐based discrimination of material stiffness and localization of nodular anomalies in simulated biological tissues. This power‐free optical interface offers an energy‐efficient platform with real‐time visual feedback for advanced haptic applications in prosthetics, medical robotics, and human‐machine interaction.