Moisture-driven, self-powered noncontact sensing interfaces via turbulence-tailored hygroelectronic effect

Noncontact human-machine interfaces (HMIs) provide a hygienic and intelligent approach for the communication between human and robots. However, they are limited by the interaction distance and bulky power supply. Here, we introduce a self-powered, noncontact intelligent sensing interface based on moisture-driven electricity generation and machine learning technique. We demonstrate that a hydrogel doped with ions exhibits strong hygroelectronic behavior and generates sustainable voltage up to ~0.6 volts from ambient air. The motion of a human hand creates localized air turbulence, resulting in changes to humidity and air pressure that tailor the electrical output. By using machine learning models to decode the motion-dependent voltage, our system achieves high gesture recognition accuracy of up to 99% for Arabic numerals, with an impressive interaction distance up to ~8 centimeters. The proposed system is demonstrated in applications such as encrypted information transmission, virtual reality gaming, and real-time vehicle control.