Self-Powered Wireless Devices and Machine Learning Enable Local Humidity Monitoring

Self-powered wireless sensors are a promising design paradigm for Internet of Things (IoT) devices that need to enable continuous, real-time environmental monitoring using energy harvesting devices. However, currently available energy harvesting devices such as the triboelectric nanogenerator (TENG), generally have either millimetre-scale environmental monitoring devices or self-powered wireless transmitters, but not both. Typical low energy density of environmental energy has limited self-powered wireless sensors' long-term operation to laboratory settings and prevented their widespread and large-scale deployment. Here we have developed a battery-free wireless moisture meter with simple structure and high accuracy. Our device is based on off-the-shelf parts suitable for various sensor and artificial intelligence applications. In the detection process, the light signal from the droplet-powered triboelectric nanogenerator is detected by the machine learning algorithm, and the humidity level detection accuracy is up to 97.3%.