Deep-learning-assisted self-powered wireless environmental monitoring system based on triboelectric nanogenerators with multiple sensing capabilities

Abstract

Since the invention of the Triboelectric Nanogenerator (TENG), researchers have used this technology to facilitate the harvesting of environmental energy, such as wind and vibration, as well as the sensing of environmental parameters, such as wind speed and geological changes, resulting in the development of self-powered environmental monitoring systems. Nonetheless, the complex and ever-changing natural environment requires that the system not only complete a wide range of monitoring tasks but also provide remote and instantaneous wireless sensing capabilities. This work proposes a self-powered wireless environmental monitoring system based on a configurable Rotary Switch TENG (RS-TENG), which integrates deep learning algorithms and is capable of multi-parameter environmental monitoring. The RS-TENG can work with wind and record speed information; a capacitive strain sensor equipped on bridges and an inductive weight sensor plated around mountains are also considered monitoring nodes for multi-parameter monitoring in the natural environment. A resonant circuit and a tip-discharge structure allow the steady transmission of wireless signals that carry significant bridge component deformation and rockfall information to the user interface for decision-making. With deep learning algorithms, the system detects the deformation states of bridge components and rockfall warning signals and identifies defined levels. By adopting a shared wind cup, the RS-TENG allows the node to power electronic devices while enabling entirely automatic wireless wind monitoring. Furthermore, a user-friendly visualization interface is built for this environmental monitoring system, which allows users to assign monitoring tasks and get reasonable results. This work provides a paradigm for TENG technology applied in self-powered wireless environmental monitoring in complicated and particular natural environments.